Javin – Reviving my desire for an A.I. player for X-Wing

I’ve been interested for a while now in Javin a projector-based visualisation tool for X-Wing, which would have several components ultimately:

1) A projector table to show off visualisation software, such as Vassal or a bespoke coded program.

2) Interactive table features to allow AR tags on ship bases to be recognised on the table as a hybrid computer/physical game interface.

3) Artificial intelligence opponents that could play either a virtual opponent or a physical opponent.

Sadly, the two students I put on this over 2013 were rather poor in delivering results, so the project was sort of shelved for a while. Anyway, I am now looking at Vassal as the main visualisation software, rather than creating our own testbed, because it has become ubiquitous in online play; the exceptional work in appearance and functionality of Vassal is well above anything that I could do in the short term. So, I am now looking at ways to modify/extend Vassal and/or the X-Wing module to allow asteroid and ship positions (x,y,ø) to be transmitted to an outside process for A.I. planning of ship movements. These would then need to be passed back to the settings for the AI. movement dials, ideally, but we need to be realistic. Many rulings and interpretations of pilot abilities and general implementation of movement would still require a human to apply.

Anyway, when (if!) I make any progress on Javin I will, of course, post it here.



I’ve been playing with Vassal basic editing, and come up with a ‘canon’ extra fighter, the TIE/d (Note, not the TIE/D which is the Defender. In this case the ‘d’ stands for ‘drone’). If you would like to run these, the important thing to do as ‘human’ players is to simulate a basic A.I. and have craft engaging in swarm (as in bot) behaviour; keeping apart, but pursuing the nearest enemy, regardless of ‘who’ is tactically best to attack. NOTE: This player should be immune to ‘Biggs’ despite the fact that this is not specifically mentioned on the card I created, as it only just occurred to me; this depends upon ‘how’ the Biggs Effect works of course. Also, on considering the weak A.I. of the lookup table (Part 1, Part 2, and Part 3) / Web Applet as a suitable model for the first version of an inbuilt A.I. that system ignores assigning Stress Tokens, because it would occasionally produce illegal moves otherwise; when a Red move came up and the ship was already stressed. There are two solutions to this:

  1. Allow the move, but stack up the stress counters, which would represent increasing confusion on the part of the bot brain and prevent actions until all had been cleared off.
  2. Update/extend the lookup table to account for occasional restrictions on movements, which would complicate the basic A.I. table quite a bit.

Pilot Card

Ship Token (with stats)

I’ll include a download link here once I’ve finished the Vassal extension and tested it.

  1. Mike this looks really cool and interesting!

    I’ve actually just finished studying a unit on Intelligent Agents at the University of Bath, so I understand what you say when you mention that you “thoroughly expect any bot to be much, much worse than an average human”!

    Is your first step setting up an Agent to play X-Wing on Vassal? I guess it will involve modifying/creating a plugin to interpret the game engine to create percepts and an environment representation for the agent can interpret and then reason about?

    From there I guess you will go down the AR route, to give you that table based version?

    Also I saw you lectured at the University of Wales, Newport. Are you based there now? I’m actually from Caerleon originally – small old world!

  2. If you can pull it off, it will be very interesting and is going to draw in so many new players in – a class of them will shy away from competitive play against humans, but some will bridge that gap (kind of like what happens in Starcraft 1/2, a lot of people never even dare try ranked games against live opponents).

    I’m about to subscribe to an A.I. class to a remote university program myself this semester too!

    Mike sent me a PM to ask if I knew a primer on the vassal code, but I’ve never even read a single line of it. I merely edited the modules (using the launcher program of vassal) after wave 2.

  3. mu0n has been very helpful 🙂 and I’ve been doing some digging into Vassal programming. The short term route plan is:
    1) mod vassal src to dump object location data to console;
    2) I eally I’d like this as an extension to X-Wing so you can selectively include it with the stable, virgin Vassal app;
    3) Have agent code inside this extension or communicating with it that can suggest moves for select ships;
    4) have suggestions realised as set movement dials in game.
    5) auto identify collisions and overlaps of move templates (rather tricky, but really useful to help novice players) so that a warning can be given to players to determine rule infringements;
    6) improve the A.I.
    NOTE: the idea here is still to require a human to get, check and implement moves.

    The medium to long term plan is to incorporate AR tags as game state input to allow physical games.

  4. I’d play the Trek version in a heartbeat if the quality of models (and size of them relative to each other) was decent, but I’ve heard only TERRIBLE things. There were some really excellent mechanics added to it though, things X-wing could have used and didn’t for whatever reason. This particular fan of Sci-Fi could never choose one side of the fence — how can you hate one and love the other, when they were both so well done?

  5. I second this enthusiastically.

    I had intended to pick up Attack Wing when it came out, and play both. However, the quality of Attack Wing is terrible. They are out of scale, poorly molded, and have atrocious paint jobs. If I were going to be happy with monocolor paint, I could buy unpainted stuff and just spraypaint it whatever color I want. The models look like crap, plain and simple. If Attack Wing were a board game that came packaged with a couple dozen of these things, I would feel differently. But it is a miniatures game, packaged and priced appropriately, and it needs much, MUCH higher quality than the junk I see on the shelves.

  6. I am heavily into X-wing, casual into Attack wing.

    One thing is Attack wing is a lot easier to play casual. 2 and 3 ships is the norm and 4 is a “swarm”. You don’t need multiples of ships (tho there are a few of the better ones where its nice if you want to play faction pure) With a core set, and 6 xpacs I went 3-0 in the last Organized Play Tournament.
    The Organized Play Tournaments are nice, tho I am not the biggest fan of exclusive gameplay elements.
    Attack Wing has a much bigger Mission-focus with a new mission in every xpac. I look forward to doing the missions with friends.

    The paint job sucks, but with some fairly minimal effort painting, they look a million times better.

    The squad building options are endless and only get bigger with a very fast production schedule.

    Attack Wing has major balance issues tho. Since releases are so quick there can be no way they adequately playtest. The ships range from very powerful to absolutely terrible.

    Most ships are 1-2 agi, tho cloak add +4. Attack values of 4 and 5 printed on the ships are common which a lot of very easy ways to increase that number. And since “buckets-o-dice” strategy is a very effective way to deal with cloak and win in general. The side-effect being that it makes the game far less tactical. Tho builds based around movement and getting out of arcs do exist. The much higher attack to defense ratio does make for shorter games.

    I still prefer X-wing mostly because it is more tactical and balanced. But Attack wing is fun.

  7. As mentioned by other posters, there is a clear difference in the quality of the miniatures, with x-wing on top. Also, and more damaging in my opinion, is the quality of the card design. The generic “federation captain” is a good example. It looks like a low resolution image files was used, the character’s head is cropped off at the top and the card’s body is dominated by a giant white box. Their is no flavor text on any cards either.

    Game play wise, I prefer X-Wing’s handling of obstacles: what happens when players fly through them, and how the game handles ships overlapping other ships or asteroids.

    Attack Wing does have some really good mechanics such as cloak, but I find I prefer the small shit dog-fight feel of x-wing to the big ship feel of Attack Wing.

    The one area I feel Attack Wing excels at is scenarios. However, I hope that the upcoming Cinematic Play will fill out this aspect for X-WIng.

  8. Attack Wings component pieces are just awful quality. Cheap feeling plastic models with atrocious paint jobs. Cards that are curling straight out of the box and low rez blurry movie stills for art.

    Gameplay is fairly decent although the balance is pretty questionable in some areas. Some factions are shafted pretty hard on captains and crew. One faction has double the amount of captain options than the other factions. You can use any upgrade on any factions ship but “pure lists” will ultimately not be very competitive for some factions. Imagine if X-wing had zero decent Rebel pilots and it let you use Howlrunner on either faction. Pretty soon every list imaginable, for both factions, is running Howlrunner in it. This is essentially the dilemma of Attack Wing.

    Their tournament scenarios can be fun but quite a few of them are just completely ignorable and turn into a straight deathmatch. I was at an organized play where every single player ignored the scenario entirely.

    I dont want to knock the game too hard as it is pretty enjoyable and they do have some fun elements. Cloaking is an interesting mechanic. Also the varying firing arcs are kinda neat.

    One thing I completely hate about this game is Warp Core Breach. This card is completely ridiculous and especially devastating when running a 2 ship build. It really swings some games too far into the luck zone more than dice ever could.

  9. As a reference point, I have 11 attack wing ships, and have only played a few games, I’ve giant x-wing fleet. I tend to fall on that both side of things…

    I think the dials in X-wing tend to represent the ships quite well. The individual ship classes have very different feel to them. The attack wing ships dials don’t always feel as separate from each other, and can feel more faction specific. Sure, the attack wing ships have some dials that feel really specific to an individual ship – but they often feel more generic and are maybe more “faction defining”.

    X-wing’s board game sized upgrade cards are really helpful for my storage solution. I really feel like the star trek cards take up a lot more space proportionally. And when we get into playing – the trek cards take up a lot more table space.

    Fantasy Flight has been noted on several occasions to have a great customer service. I’ve had replaced parts shipped to me in japan with no charge. I’ve seen a story of them replacing someone’s cards/punchboards who lost them, and dials that had been destroyed. (I don’t have experience with wizkids,)

    I feel X-wing is less dependent on luck.
    for example – The focus action as a core game mechanic in X-wing (Every ship thus far has it) is instrumental in affecting the probability of the dice. Comparatively, Battlestations in attack wing however, is absent on many ships – giving us less control over probability when using those ships. (so it becomes part of list building in trek) I feel this makes attack wing more random and “swingy” than x-wing.

    Also attack wing often has combinations belting out higher amounts of dice, (I think we’ve seen as high as 9). – The added dice also contribute to more variance in results.) X-wing maxes out at 6 – and you have more control over how that works with focus an option)

    X-wing appears to be responding to the tournament meta. The upcoming imperial aces pack seems very much to be trying to infuse more interceptors into the game – the falcon’s turret has been a pretty big death sentence for interceptors. The previewed interceptor pilots and upgrade abilities seem to be a response to this, and a “re-balancing” of the interceptor.

    and one special case… X-wing’s punch boards come with multiple “generic” ship tiles for each ship. (due to the pilot system) – Attack wing has a single dual sided ship tile. Since attack wing re-used the klingon ship mold for the Kronos one and the tong, if there had been multiple “tiles” it would’ve been possible to use the models as either the k’tinga or the d7 – a bummer for tournament play. While the new interceptors will be able to use old interceptor tiles, which is pretty cool.

  10. Now let’s discuss the things X-Wing doesn’t do so well at, that Attack Wing makes it so! Three plus factions would make for more varied play, whereas it’s either Empire vs Rebels or mirror matches at present. Missions in each AW expansion are well rounded and those in XW are rather poor. AW has a generous number of cards in each box, where FFG requires you to buy Lambdas to get those B-Wings their lovely Advanced Sensors cards, etc. And I have to wait for a Luke card to allow him to fly anything other than an X-Wing, but AW can have a Ferengi crew the Enterprise because captain and crew cards can be allocated to any ship. Finally, AW gets reverse manoeuvres, and the Cloak/sensor echo and other mechanics would be nice for XW wouldn’t they?


  11. Don’t the faction cards just reflect the canon and lore of the Star Trek universe though? And would XW players not enjoy scenario play at competitions? What would you change about XW firing arcs, when they seem pretty representative of Star Wars?

  12. Some really good points. Interesting you say AW is more luck based when craft are more likely to hit and more resilient, which favours slow attrition rather than quick (lucky?) kills. The meta tournament tweeks of XW might still come to AW though as it’s younger. And you can’t blame Wizkids for fewer ship tokens where players could ‘repurpose’ ships from previous games otherwise…

  13. I recall a BGG thread where the same was said of the core set, if you could see the rough edges of the dials where they’d been attached to the card stock. However, it does seem silly to have put a big image on the back. Who know, maybe it’s more thematic to realise Kahn is guilty of two dimensional thinking.

  14. In Attack Wing, you can shoot at ships that you are touching, whereas X-wing you can not. Its a small but significant change that decreases tactics. Other than that it is still the same where you lose your action and take damage from obstacles.
    Another very small change is obstacles in AW are circles whereas in xwing they are varying shapes.

  15. I agree that Attack wing has the edge in squad building options (but whether it has as many top-tier competitive squad building options remains to be seen. Also some of these top tier builds require exclusive cards from winning OP tournaments) and has the edge with missions (tho honestly AW can keep that niche imo)

    Reverse actions, 6-forwards, 4 banks, and 180 arcs (tho if we get the Outrider, it will likely have one) are fun, but I don’t feel like X-wing is missing something by not having them and that stuff isn’t too thematic in this universe anyway.

    As for actions, I think the effect that cloak has on the AW metagame is a detrimental one, it renders a ton of cards near-useless. Hell, it completely nullifies first engagement round torpedo alpha strikes…from a design perspective, how is that a good thing? I quite like scan – by itself its not great, but with card investment its pretty good.
    X-wing has the boost action and barrel roll action (which isn’t reliant on cloak like in AW) also promote more “player flying skill” in X-wing.

  16. Technically, just FYI, high numbers of dice reduce variance, not increase it. “Variance” is distance from the mean.

    On 1d6, a result of 1 — as far from the mean of 3.5 as you can get — is just as likely as a result of 3. This is high variance. On 2d6, a result of 2 — as far from the mean of 7 as you can get — occurs only one-sixth as often as a result of 7.

    The probability of proximity of actual result to mean — when compared to the range of possible results — increases as dice are added. Thus variance decreases as dice are added.

  17. I thought you said in OP you were chosen to argue for X-Wing side … This post, plus other comments makes it seem like you are pro Attack Wing?!? Is this just you trying to be open minded, “knowing your enemy”, or is this coming debate rigged?

  18. You have to pay a 1 point penalty for every upgrade you put on a ship that doesn’t match that ships faction. Some upgrades are also faction exclusive.

    I think XW players would enjoy scenarios if they were actually interesting. Attack Wing scenarios are either boring (some space cannons shoot you) or ignorable (shoot this thing for a trivial amount of points while your opponent ignores it and kills you.)

    I wouldn’t change the XW firing arcs. Each game has arcs that represent their respective ships (fighters vs larger battleships) pretty well.

  19. Though – you can see an increasing linear trend in variance as you add more and more attack dice –


    just note that it’s comparing focus vs expose at a set number of dice – so the variance for expose at 1 die – is the variance on 2 dice (since expose adds 1 die to the result – this was shifted as such to more easily compare the variance between the choice in actions on a ship with a particular attack value – sadly, i didn’t keep going for the 9-10 possible die values in attack wing.

    focus being battlestations… many ships don’t have battlestations …

  20. I’m guessing this was due to a really well compiled captain and crew list, resulting in an obscenely large number of hit dice. While it’s good to know that AW can be a bit more chaotic, this does seem less thematic. I recall, from TOS and (less so) the movies and TNG, cloaking being much more what I would call ‘hidden movement’ and Inwas hoping that sensor echo would be much more like an extra full movement than a ‘sort of’ barrel roll.

    “Captain, we are detecting what ve believe to be a cloaked weasel off the port bow!”

    [Bird of Prey decloakes directly behind then starboard nacelle…


  21. Movable pilots would probably be nice – though, I think the fixed pilots help you to better control game balance. Attack wing has to keep on it’s toes not to make broken pilot + ship ability combinations. But – other than switching the ship type – I think really the whole mechanic of adding titles gives you the “named vs unnamed” ship thing – so at least there’s that.

    I’m not sure that attack wing actually has the edge in squad building – consider the possibility that there is one “best” build for attack wing, that list being the same, while x-wing will have “two” best builds as the factions can’t be mixed. Of course, in truth, you can play faction pure, and then you increase your variation a lot in star trek.

    Attack wing currently has – Tech, Weapons, crew, talents and captains right (And in a way titles)? X-wing has modifications, titles, weapons, systems, crew and pilots…. so it’s eeking out one more category (I grouped all the weapons stuff together) – you could group all of the systems/droids as tech I suppose), but the actual number of icons in X-wing is larger, meaning there’s more tactical variation across the field, a larger design area if you will… (X wing also has that new epic scale coming into play)

    Sensor echo might be cool, but, the cloak as shields, I don’t think would function so well in X-wing (Given that all of the ships have the focus action) – meaning you would be looking at what – 7 agility dice with focus for a cloaking tie fighter – good luck hitting that thing…

    And – even though AW is generous in it’s cards – you still have the “Buy X if you want the Y upgrade”

    As for things I think attack wing does better at the moment?

    +Multiple factions

    +OP events

    +Mission with every expansion

    +Terrain / mission tokens (Space station, mine field, scimitar, planet etc)

    +Overall number of ships – (in a game about ship combat, that’s got to be a plus)

    + Future roadmap a bit more laid out (future waves) – who knows what’s coming down the longer road in x-wing.

    +Better quality rulebook (somehow, while the card quality isn’t so great – the rulebook feels pretty nice to me)

    +Public Presence – Andrew Parks is pretty active in the BGG forums, giving answers, advice and hints of spoilers… FFG tends to stay pretty quiet

    +Weapon pictures art – It’s pretty cool that each photon torpedo is matched to the ship that it comes with – even if they are the same “stats” – all of the x-wing proton torpedo’s have the same art. Some people might prefer the sameness – but I like the feel of the variation.

  22. it was 4 dice – vs the cloaked maht-ha which is 5 hull (no shields but lots of agility dice). Range 2

    it was all hits, and one or two were crits, he rolled 1 evade on 5 dice – and Geordi lets you force them to reroll an evade die. with the direct hit – Kaboom, gone.

    funny thing is that something like this happened in 2 games in a row…

  23. There’s a spectrum of possibilities. To claim that you need to spend $400 on X-Wing to play without constraints is a disservice to your argument.

    If I had gone the sub $70 rebel route, then I’d necessarily have access to imperial players (or “omni-players”) from whom I could borrow cards. I’m merely pointing that it’s a possibility if someone wants to be frugal.

  24. I agree. Remember it wasn’t me that claimed $400 was necessary. Taking mu0n’s point at face value, he was (I believe) making a more neutral statement about the fact that most people will have to decide one way or another, due to cost of expansions. What is good about XW from your comment – and it is an important point – is that there is no “pay to win/take part” curse, compared with some games. This is one of the benefits of the XW community, I think, because of the “fly casual” have fun attitude that most of the players show.

  25. More pilot movement (probably only for uniques), but maintaining game balance, could be possible with card only expansions. 🙂

    You make a good point regarding the range of upgrades, but remember that AW has 4-5 slots per ship, which is a lot more than XW. However, you are right about the “Buy X to get Y…” issue.

    Cloaks are more temporary Agility, rather than Shields, I think. However, even with 7 agility dice it is still possible to have a bad roll. As to the physical components that AW has uniquely, I think that mission packs, with tokens and better written scenarios might again address this; FFG if you are reading this, some of us would be more than happy to write these 🙂 Art for the XW upgrades could be more differentiated, while not being ship specific, but at least it is all bespoke commissioned work, rather than recycled screen grabs.

    Jay Little is quite proactive in public support of XW, but I like especially the fact that OP events run/sponsored by FFG only preview the next wave, with few (if any) unique prizes and giveaways that won’t come (in some form) to the wider player community at some point.

  26. Cloak doesnt just give 4 defense dice, it also blocks target lock which prevents both rerolls and torpedoes. This has an effect on the ships/cards that people bring to the table. And as a result of cloak, many cards are not worth bringing because they are weak to it. (Ships with low attack values, torpedoes, Gor Portas would be an amazing ship if not for cloaks presence)

  27. It depends on what you mean by “luck.” Reduced variance (e.g., more dice) will result in fewer rolls that fall far from the mean. So in that respect, the influence of luck is lessened.

    But when those extreme rolls *do* happen, they have an exaggerated impact. So in that respect, the influence of luck is increased.

    By reducing variance, you’re making lucky single events rarer, but more impactful when they happen. E.g., rolling a 1 on 1d6 is barely even unfortunate; rolling 3 on 3d6 is disastrous.

  28. True – probably half a dozen of one, 6 of the other when it comes to the upgrade slots…

    I agree that Jay is pro-active in pubic support, it’s just that you don’t see FFG giving answers in the rules section of BGG – they seem to stay back and observe, eventually adressing things via an FAQ – or individually answering through the FFG support site contact form.

  29. Doing my research 🙂 Jay is MUCH more active on BGG than Andrew; info from their respective BGG profiles. However, you make a good point about FFG on BGG forums, but they do run their own forums for X-Wing, which are incredibly useful. Not sure about WizKids, but it is clear than Andrew (by an earlier account) is more public on BGG, so maybe they don’t need their own forums?

  30. I have posted my thoughts on the two games in my blog:
    X-Wing vs. Attack Wing | IMMATERIUM PRESS

    Some of the ratings I could adjust (dropping STAW scores slightly due to balance and more manufacturing issues) but it’s still a pretty good game.

    Overall I prefer X-wing but STAW does have tons of organized play support, which means I play a lot of games even if it’s not my #1 choice.

  31. My concern with reliance on OP to popularise AW is that it will fizzle out, if only because of the (too?) quick release schedule. WizKids are riding a wave (sic) right now, with ‘exclusives’ baiting the trap; emotive language I know, but I can’t see this level of PR (I hesitate to use the word ‘support’.) being maintainable financially. Once sales dip, which inevitable they will do with diminishing returns, my worry would be the fast track approach will peter out.

    FFG has a slower release schedule because they have to make new ships, balance new cards and upgrades with previous ones, etc, but this modest pace may benefit XW by making for a more mature game in the long run.

  32. They have another OP lined up that will get a lot of attention – the borg. I know multiple guys who are going to jump in once the borg go live, and this should sustain high interest for the rest of 2014.

    While X-wing is more balanced and tournament friendly, the OP system for Attack Wing definitely drives up excitement and keeps people in the stores much more frequently than with X-wing (from what I can see regionally, anyway).

    I think both systems have legs, and should be around for another couple years at least. A lot of it will depend on how they are able to incorporate the products from the new movies, as they will surely run out of credible releases from the old background material sooner or later.

  33. I think that reinforces my point. OP based on exclusives panders to the fan boy, but it’s mercenary in its bought loyalty. Without the OP hype would it fade quicker into obscurity. If decent tournament play continues, with a few new shiny things released every so often, I imagine a long tail for these games. Otherwise they will fade fast.

  34. There are several responses and maybe this has been said before but the biggest thing that gets under my skin about attack wing is that you can put any upgrade or captain on any ship just by paying an extra 1 point for them. Therefore it creates an “Ultimate Strategy Squad”.

    Also staying competitive seems tough with Wizkids releasing new ships every month. The cost of playing Attack Wing is getting ridiculous very quickly.

    What I do like about the Attack wing is that the games doesn’t take as long because the high attack to next to nothing defense rolls.

    Sure I wish they would make better models, however I would be happy if they just made them cheaper overall. It is still a fun game but I think it just can’t hold it’s own when compared to X-Wing.

  35. … the biggest thing that gets under my skin about attack wing is that you can put any upgrade or captain on any ship just by paying an extra 1 point for them. Therefore it creates an “Ultimate Strategy Squad”…

    You hear this a lot, but a LOT of areas play “faction pure” lists to avoid shenanigans. Even a lot of us hard-line folks are softening a little on this stance, saying that each ship in your fleet should be faction pure but for example adding a Vulcan ship when they come out this year won’t ruin Federation.

    …Also staying competitive seems tough with Wizkids releasing new ships every month. The cost of playing Attack Wing is getting ridiculous very quickly…

    I find Attack Wing is actually a lot cheaper. If you collect one faction, you don’t have much reason to take more than 1 of each ship. There’s not as much incentive to take multiples as there is in X-wing (and most of the good upgrade cards are unique anyway).

    …What I do like about the Attack wing is that the games doesn’t take as long because the high attack to next to nothing defense rolls…

    I’ve actually never had a timed X-wing game go the distance. Happens a fair bit in Attack Wing since ships are tough and many heavy hitters can’t turn very well (no k-turns). If you survive the alpha strike, it’s hard to finish them off.

    …Sure I wish they would make better models, however I would be happy if they just made them cheaper overall. It is still a fun game but I think it just can’t hold it’s own when compared to X-Wing…

    Amen to that, it’s hard to pay the same price when the manufacturing quality isn’t as good. Still, I’ve paid maybe $100 total for all my purchased Attack Wing (starter + 4 fed ships). I don’t like it as much but I’ve spent almost as much money on Lambda shuttles as I have on all my STAW!

  36. Well, I’ve had confirmation that I have been chosen to defend XW in the Dice Tower Showdown debate; we auditioned in the early hours of Tuesday morning (1am GMT, 8pm EST), but a place wasn’t certain as there was a lot of interest in participating. The show will be recorded live next Monday night/Tuesday morning, with few if any edits/cuts, and I’m guessing will be out quite soon after that.

    You’ve all done a great job in providing the pros and cons of both XW and AW. If we ‘lose’ it won’t be down to poor prep. However, don’t forget to vote once the show airs, as with any debate it’s the end vote that counts 🙂

    I’ll post a link when I get now for the show.

  37. Dice Tower Showdown: X-Wing vs Star Trek Attack Wing podcast is now live:
    Listen to it then go here to vote (It’s near the bottom of the page, ep13 and needs BGG registration to vote):
    Personally, I thought we (XW) had the better arguments and the AW defenders were rather objectionable, with constant interruptions…

  38. I agree with the constant interruptions, the AW defenders were very argumentative and did their best to keep you from getting your points across, I’m impressed you managed to not lose your temper!

    The AW guys spent far too much time arguing that X-Wing was bad rather than singing the merits of Attack Wing. Your Dogfight vs Boxing match analogy summed it up for me.

  39. All I heard was “they have the same mechanics and AW has 42 ships announced so it has more variety, except when you get into factions which have 5 ships (their example).” So less variety by faction??

    Personally I enjoy Star Wars more, though I did take one look at the Attack Wing models and decided that they were a waste of my money.

  40. I’ve played in two Store Championships, with a total of 31 entrants, and there was only one seven-TIE swarm. (There were two other six-TIE swarms.)

    I don’t think you should’ve conceded the “everybody flies TIE swarms,” even when only considering tournaments.

  41. I agree, but it was hard to avoid the constant barrage of “fewer ships, TIE swarm=broken” because that’s pretty much all The AW side said; actually, shouted, constantly, interrupting us…
    We did try to address that criticisms couple of times, but maybe it could have been better defended. Please remember though that this was recorded quite a while ago. Alternatives to swarms are, by our community, much more recognised though.

    However, you should post your sentiments on the BGG page above, where there are lots of comments regarding the debate.

  42. 3 B wings, heavy laser, fire control. They will destroy sitting ducks. Once in range you move one forward and barrel roll back to keep in range. Being at range 3 will give you the advantage of an extra evade dice. While your heavy laser w/ target locks on these 3 ships will average over 9 hits. And a good chance at landing a crit.

  43. I mean, any group of ships can take this out. you’d skirt the edge of the map and engage at the arc of the rearmost lambda so only one shuttle can shoot you and focus that rear lambda. then, if you don’t kill it, you ram all your ships into the rear lambda and focus fire the other ships that are just chilling out in front of your guns. It’d be pretty easy for nearly any squad.

    I’d think it’d be a little harder with falcons, but I’d think if you’d like a lot of range 1 shots on one falcon, just head all your ships to ram into the non-target falcon so it can’t shoot any of your ships and then fire all range 1 at the other falcon.

  44. Correct about APL only triggering if you end up touching the big ship. I used my own blues as shock absorbers once or twice to avoid APL hits last night en route to a store championship win 🙂 Eventually did ram the stupid ORS but not before lighting it up twice at range 1 with TL/F, allowing Chewie to finish it off before it really became a problem.

  45. I just found this interesting interview with a professional poker player:
    which discusses the fact that aggressive poker players do increase the randomness of games. And another discussion of Hindsight Bias:
    This wonderful piece on the value of morale seems quite relevant:

  46. Paul did not have a first round bye, quite sure of this, as he has never won a feeder tournament (a Nationals) to Worlds.

    The Challonge bracket is correct. Look again at the values presented. There is no way they could represent points destroyed, they are prediction percentages based on votes casted on who would win the given match-up. Ie. – Paul and Morgan were 50-50 on winning according to people’s predictions.

  47. X-wing is a game of skill+luck, not just a game of skill. This was Sunny’s opinion, and it also seems to be yours. But where Sunny says that X-wing is mostly skill, with an element of luck… you seem to be arguing that X-wing is mostly a game of luck, with an element of skill. Or if I am generous in reading this article, you seem to be arguing Sunny went too far in saying that skill is more important that luck when playing X-wing… that it’s not demonstrated which is more important.

    I would say that Sunny is correct, and that you are in denial of evidence.

    If the game is primarily luck with -some- skill involved, you have no more than a lottery chance of a repeat win. You will not see the same names come close to winning repeatedly, over the course of time. You will not see the same names win or nearly win several times in a short window of time.

    But look at my track record. I have qualified for all but one (my x4 Y-wing list in HiLo went 2-3) Vassal elimination since we started having tournaments here at TC. I don’t believe I am the only player with that claim. Paul and Dom BOTH have successful back-to-back title defenses. By these things we know that X-wing is not primarily a game of luck — not by any one event, but by a pattern of similar events. We would not see this pattern were luck a large determiner of outcome. And we would not see the same players struggle either; they would sometimes do incredibly well.

    X-wing is proven as a game of skill — exactly as Sunny professes.

    I want to break the game down here, with reference to luck:

    A number of of games are pretty even. I’m sure most will agree that in these games skill is a factor, that the better player will win more than 50% of the time. Perhaps how 2 squads match up also matters, but I would actually call bringing a solid squad a skill so we are still not speaking much about luck.

    However, in other games one player or the other is favored heavily. Abstractly, any player will see a 50/50 split — favored sometimes, fighting bad dice other times. If luck is the primary determiner, then all players will have similar W:L ratios when seeing uneven dice. But we don’t see that.

    Instead we see those same players whose names do well most often, winning games when dice favored their opponent. We see the same players manage to lose sometimes, even when the dice favored them heavily. Skill continues to matter tremendously.

    I would say that every player has 3 ratios, determiners of their skill:
    -1- W:L record in even games
    -2- W:L record when dice favor them
    -3- W:L record when dice favor the opponent

    The argument of players with a poor 3rd ratio is that X-wing is sometimes just a dice game, that they were doomed and there was nothing they could do. For a portion of those losses, it’s true — but for most of those losses, they are just making excuses. Another player could take those same rolls and win the game with them. Instead of admitting to poor play, players are professing a flaw in X-wing that it does not possess.

    It is this denial of reality that Sunny argues is largely put to rest by yet another big example of the same player doing well, because it came on a very public stage. Worlds is the biggest tournament, and well advertised. Those holding that X-wing is just a game of luck took a major body blow, and I agree that this argument is pretty much dead now just as Sunny says.

    I would estimate, inductively by my own experience playing X-wing, that fewer than 1 in 6 games of X-wing are determined primarily by the dice. Players are lucky in 50% of those games, so bad games come roughly 1 in 12 times you play — and even in these games, skill can steal wins if you outplay the opponent well enough. The other 5 are determined by relative skill, with luck playing a part in those games to varying small degrees.

    I want to state separately that in a vacuum, Paul winning World twice would not mean anything. With no competition history for X-wing available for reference, it would be entirely reasonable to think that his feat might be mostly luck… that he played well enough to get shots, and just rolled his way to victory twice. But we are not in a vacuum. Paul’s win is not the coffin, just the final nail. You have to play well to win.

    Another aspect of this article that bothers me is that there’s a quiet assumption assumed, that players who win had to be lucky to get there and be ultimately victorious. It is entirely possible for a player to see nothing but bad dice and still win game after game, by playing well enough.

  48. Yes and no; Skill does not mean “potential skill” or even “average displayed skill”, but really performance specific to that game. I would expect that in a Paul Heaver vs Paul Heaver game, the side he played better would win… although dice might flip the outcome if skill was equal, or create a more severe outcome.

    It’s definitely Skill + Luck. I don’t think too many people could argue otherwise, in any serious way. The crux of argument is how much these things impact who wins relative to the other… in a way, what the ratio is between Skill and Luck.

    Put most simply… how much better do you have to play to make up for a bad roll, and how much worse do you have to play to marginalize a great roll? For any game these questions have a real answer, and in X-wing any single small decision can sometimes erase SEVERAL rolls that went one player’s way. The ratio majorly favors Skill over Luck. I say it’s roughly 6:1, conservatively 4:1 at least.

  49. I concur with Theorist’s conclusions and I strongly disagree with a lot of this article.

    I appreciate that it is thought provoking, but I agree with Theorist that a lot of the argument and conclusions are misguided.

    If I had more time I’d go through it line by line and dissect.

  50. I appreciate the candor of the article and I take no offense whatsoever to your opinions, nor your disagreement of mine. I welcome the conversation.

    However, I think that you’re extrapolating my point to an extreme and nearing the point of putting words in my mouth when you say that I would
    1: attribute no luck to winners
    2: write the same article if someone else won because they “won all their games”

    On the first point, the article’s main drive is the following statement; too many players attribute a larger percentage of luck to xwing than is reality. Of course X-wing has an “element” of luck; it has dice. There are lots and lots and lots of dice rolls. However, the element of skill involved is, in my opinion, capable of nearly completely overshadowing the luck element. I would say that if I had to guess, I could fairly look at the results of the top 32 qualifiers, look at their resumes, and say that it’s about 90% skill that got them there versus those that didn’t.

    The second point is simply wrong. I would not have written this article if Bob from my local game that rarely wins had won Worlds. I think that would go to showing the opposite.

    You go to lengths to defend your own play as “I entered it to have fun”, which I find bizarre in the context of the argument. Are you implying that the game is so luck based that you no longer feel the value in competing?

    In your point regarding fortressing ships, you imply that Heaver flying a Fat Han is a decision was a win at all costs strategy. This contradicts your statement about your own experience when you do “try” to win: “It has certainly helped me in facing regular defeat, except when I fly TIE Swarms :-)” So if Fat Han is a win at all cost strategy, why wasn’t the top 8 all Fat Han’s? Or was that simply because there is too much luck to allow them all to defeat “inferior” lists.

    The poker argument is a subtle one and not without mistakes on my part. The line from Rounders is innacurate in todays Main Event at the world series, which is by far the most viewed event. It is not, however, wrong when looked at in the big picture. I had listened to the Radiolab conversation which was largely a conversation on how to play poker calculating the unknown and I don’t think is quite as applicable as you’re placing it.

    The Main Event is so gigantic that given the luck/skill ratio of poker, in a single event, the skill versus the sheer number of opponents makes it very difficult to get to the final table. Contrast that with 1: other poker games (HORSE, Stud, Razz, Omaha Hi/Lo) and 2: nearly any other size event, and you will see more realistic results. However, in each of those events, you will see a different skill/luck ratio. Great players do better in HORSE tournaments than they do in Razz. I will argue that X-wing’s skill/luck ratio is heavier on skill than nearly any poker game.

    The argument for luck IS being made. It is being made every day by new players who lose a game and fail to understand WHY they lost. They didn’t lose because they couldn’t roll green dice (as you imply “LOADS of green dice as my particular bugbear”; emphasis added). It’s because they put themselves in a position to have to roll LOADS of green dice in the first place. This is a key concept that perhaps I wasn’t clear enough on so I’ll give it it’s own paragraph again:

    Two players that roll equal numbers of dice are going to have their game decided by a high percentage of luck. Great players put themselves in positions where they are rolling more attack dice and less defensive dice than you. In short, they’re putting themselves in as optimal a position as possible by removing as much luck from the game as possible.

    The crux of my article (beating the dead horse) is when you think that luck is not in your favor, reevaluate your game to see why you allowed luck to beat you and if you could have avoided it. That is how you grow.

    I still strongly believe that having the same world champion twice in two years IS a solid confirmation of my point. No, he did not “beat 250” individual people. Yes, there is an element of luck. But look at the top 4, of a tournament that size, and with a straight face tell me that luck got them there. Skill is overwhelmingly the reason.

  51. Hooray, I won the comments with an Obi-wan quote! I’ll throw in another, more cringe-worthy, one: “Only a Sith deals in absolutes.”

    The absurdity of youngish Obi-wan’s statement aside, I think this is a well-written article with a slightly flawed premise. Sunny’s original article wasn’t pushing a false binary.

    “But the biggest reason why Mr. Heaver repeating at worlds is great for Xwing is this: It proves without doubt that skill overcomes dice in the long run.

    There are many games out there that suffer from the extremes on both ends of the luck spectrum… Why is texas holdem more popular than 5-card stud and gin rummy? Because 5 card stud is too much luck and gin rummy is not enough. You need a good balance for a great game. I believe that X-wing fits nicely into this slot.”

    He’s not saying dice are irrelevant, he’s saying that the game is (to a large degree) determined by skill, and that the significance of luck is simply in the goldilocks zone; enough to keep the games entertaining but not enough to diminish the ability of skilled players to perform at a constently high level over time.

  52. I defeated the Internet once. But the number of apologies and concession replies crashed everything, and by the time things rebooted people had new things to argue about. Still, for but one moment things were glorious.

  53. Luck is high enough that the best player doesn’t always win, but luck is also low enough that the best player is the favorite. That’s X-wing for sure. I wish I was as concise with my words as you, man.

  54. Paul H. can’t play better than himself. I like how you frame the argument on a scale. I also agree that it is specific or relative to each game. I like your scale too.

    That scale will change based on relative skill of the players in the match, with luck playing a bigger role as the skill gets better. Even if it’s not a whole lot, it is something to recognize.

    The best players will rise to the top by skill, but that as it gets more competitve, the advantage given by certain dice rolls will matter more, because theortically fewer mistakes are made. If fewer mistakes are made, then the relative advantage of good decisions gets smaller, since both players are making good moves.

    Some rolls can’t be made up for. Ships do get blown up.

    Paul got there on skill and won on skill, but as everyone gets better Paul may need more luck in the future. That is basically my argument. And who knows when that point is reached. Probably not there yet, but if more people practice like Paul, it could be soon.

    This is an interesting discussion. Another one might be why Xwing gets these kinds of discussions. I don’t see luck vs skill talk much in card games around here.

    Here’s a quote from my source which was an interview with a guy who studies this stuff::


    I think the critical distinction is between absolute and relative performance. In field after field, we have seen absolute performance improve. For example, in sports that measure performance using a clock—including swimming, running, and crew—athletes today are much faster than they were in the past and will continue to improve up to the point of human physiological limits. A similar process is happening in business, where the quality and reliability of products has increased steadily over time.

    But where there’s competition, it’s not absolute performance we care about but relative performance. This point can be confusing. For example, the analysis shows that baseball has a lot of randomness, which doesn’t seem to square with the fact that hitting a 95-mile-an-hour fastball is one of the hardest things to do in any sport. Naturally, there is tremendous skill in hitting a fastball, just as there is tremendous skill in throwing a fastball. The key is that as pitchers and hitters improve, they improve in rough lockstep, offsetting one another. The absolute improvement is obscured by the relative parity.

    This leads to one of the points that I think is most counter to intuition. As skill increases, it tends to become more uniform across the population. Provided that the contribution of luck remains stable, you get a case where increases in skill lead to luck being a bigger contributor to outcomes. That’s the paradox of skill.


  55. Having talked to Pual, and seen Paul in action, I can say that he does everything he can to take luck out of the picture.

    Squad Design: Paul took time looking at match ups and made choices that would fit what he expected to see at worlds. Lets look at the Upgrades that he picked. Predator, reduces the odds of bad rolls. C-3PO makes for a guaranteed evade. The Falcon Title. R2-D2 Lets him regenerate a shield. With R2-D2 there’s some risk, but Paul felt the benefits out weighed the risks..

    Prep: Paul practiced against what he saw as the main threats at worlds, and developed solid strategies to win against them. These plans were tested and refined. No luck here

    Playing at Worlds: Before the Game begins, you have the asteroid setup. This is what Paul calls “Turn Zero”. He’s written a few posts on this. Setting up asteroids and placing your ships is all about Skill. seeing what lines are available and best approaches is 100% Skill. Picking the better position in the beginning can give you a huge advantage in the game.

    If all you do is Joust and roll dice, then your giving into Luck. Paul’s Han doesn’t joust. He arc dodges and uses his defensive upgrades and high health to out last you.

    Paul did a great job with Squad design, and predicted the meta accurately. He took the time to develop solid strategies to counter his weaknesses. He then executed his plans. He did everything you can to limit how dice can effect the game, so his winning has way more to do with Skill then luck.

  56. “he does everything he can to take luck out of the picture”

    This 100%.

    I liked the bit in Paul’s article about how they put together a powerpoint with details about the expected metagame and what was the best list in each category.

    Just like anything else: if you want to do well, practice (effectively). Just like throwing a ball against a garage door won’t make you a good big league pitcher: playing a squad in X-Wing a few isolated times won’t make you a world champ.

  57. Hello Theorist. Firstly, it is important to state that I do believe that X-Wing is a game of skill more than luck, and can agree to a limited extent with the intent of Sunny’s article. I just cannot have the same confidence in his conclusions as you, because the results aren’t statistically significant. This is why I drew the analogy to the odds of someone winning in a completely chance driven tournament; not to show that Paul won by blind luck, just that him winning isn’t sufficient evidence to disprove that randomness is a factor. This is the origin of my claim of hindsight bias; we cannot truly separate out the decisions we made or the dice we rolled from the conclusion.

    The main contention here is not the continuum between “Luck not Skill” and “Skill not Luck”, but I think that we can agree that the game is a combination of both, like pretty much any sport. Ironically, it is the very effect that bad luck has that makes the most skillful players, because they regularly overcome random elements. In my piece, I do go on to discuss how ‘incomplete knowledge’ contributes to the most deterministic games, where randomness plays no part at all; Sunny did no favours arguing that Chess, for example, “sucked”. Where we may disagree is where the line is drawn between luck and skill; not because I want to argue that luck has more of an influence than you would be comfortable with, but that attempting to define where the line should be put is arbitrary. We just don’t have enough statistical data. There is no coffin, and no nail. There may never be.

    So, I was not arguing that X-Wing is primarily about luck, and your “generous” interpretation is correct: Sunny went too far in his conclusions in my opinion. Furthermore, you and Sunny both reference some mythical figures who have been arguing that X-Wing is just a game of luck; in fact you also paint me with those colours too. Who are the people who took a “major body blow” from Paul’s second victory? I did a little searching, but Googling didn’t come up with anything relevant. The only article I can find on Team Covenant about luck over skill in X-Wing was Piqsid’s first post, “Skill vs Luck. Who’s winning?”
    which seems to be more about the possibility of sore losers blaming dice. Is this what you both meant? If so, I cannot see a credible movement that needs convincing. Grumblers got to grumble. Neither Sunny, you or I will ever convince them otherwise. And yes, I do believe that Piqsquid hits the (coffin?) nail squarely on the head, by saying it is disingenuous for losers to blame bad luck, and for winners not to.

    I did pretty badly (win wise) in the tournament last weekend – my intent was to see just what I could do with my accidentally random list, and I did improve over the five games, gradually increasing the number of enemy ships destroyed and friendlies saved. So, in the last game, where my opponent acknowledged that I had some pretty poor dice rolls, I politely brushed him off with comments that this is what dice do sometimes. He was maybe trying to let me off the hook, but I refused to let him. I was not flying optimally, with a flawed list. He deserved to win.

    As for lottery wins, we actually do see people winning twice, and we do see them winning within a close space of time. We also hear about people struck twice by lightning. Million to one shots do happen every day if the shot is repeated millions of times, but let’s not get too diverted into people’s poor understanding of probabilities. I believe that you are right in your suggestion that one player with the same dice would probably win where another would lose. Unfortunately, this is unprovable, unless we pre-rolled dice results and forced many players to use these predetermined rolls. Maybe if we forced players to use the same lists and swap sides, we could begin to reduce the variables; even then I’m not sure that this would be a valid experiment. The game is just too ‘noisy’ to know, but then that is part of its depth and appeal. I hardly think it is fair to dismiss my viewpoint as a “denial of reality”. Nor has my own article seen fit to accuse or besmirch anyone here. It’s possible that your rhetoric is based on the false assumption that I was arguing that the game is dominated by chance, but maybe we should start by asking for clarification, rather than making assumptions. [I’ll address the concern Sunny has of me misrepresenting his arguments in my reply to him below.]

    Your inductive reasoning about your experience of X-Wing is interesting, if only episodic in nature, but I can only agree that my own experience of similar games leads me to believe that X-Wing gives opportunities for complex and emergent choices that have a major effect on the outcome. If I didn’t believe this, I wouldn’t have invested in ships, nor started to analyse the game as part of my academic research, and I would certainly not have considered it a suitable platform for exploring Artificial Intelligence (A.I.). In consulting Jay Little about the project, we held a consensus that some parts of the game are extremely complex, and might be almost impossible for an AI to grasp. The variables in a simple game also go way beyond human capabilities as well, however, and it is reading the opponent, rather than determining all possible combinations of moves, that the skilled player exploits; all players are similarly limited, so the number of permutations is reduced to a manageable level. Jay initially suggested that X-Wing would be a “hard problem to solve”, and admitted that he never even considered the possibility. [There are no plans for an X-Wing iPad App, I’m assuming.] However, after we had discussed it for a while at the UKGE 2013, he came round a bit, and conceded that while some parts of the game, such as list planning, might be forever beyond the reach of a computer, certain core elements would be interesting to explore with a virtual player.

    Sorry for the digression. I just wanted to reaffirm that I don’t believe X-Wing to be just luck. I know I am currently not the best player – too little time to play to be honest – but the article above is not an attack on the game we all love. We can totally agree that Paul Heaver deserved to win the title, because he put in a lot of effort to get himself there. Personally, I’d have loved Keith Wilson to have qualified for the final. That would have been a great match up. If I have presented any offence, such as the implication that Paul got there by luck, I am sorry. It was not the intent, but reading through the piece again, I do not see any evidence of this.

  58. While a lottery winner can win the lottery a second time, you are talking about chance events — winning 2 major tournaments at a skill-based game is not an “odds of winning” event. There is, at best, some odds that dice for that player are so bad they can’t win. The remaining chances require that player to play well. Luck is NOT required to win a game or even a whole tournament, it just helps to have some. You have inserted the assumption that luck is necessary to success as premise, which leads you to the conclusion that we can’t know how much luck is a factor. We very much can know; it is definitely measurable.

    Specifically to lightning, each time you are struck you become more polarized. Someone who has been struck before has slightly better odds of being struck than someone who never has. They attract the bolt.

    Perhaps a better comparison you might make is to juggling. Anyone can hurl stuff in the air, and with luck will catch at least some of it. But to juggle well, you have to practice and have a good technique. Juggling is skill based, but not exact — just like X-wing. There are factors outside our control, which we can minimize but never eliminate.

    You are in the unfortunate position of arguing that it’s possible for someone who has never juggled to manage 3 bowling pins over 2 minutes without dropping anything. By statistical probability, someone COULD by sheer dumb luck manage it. But that’s not reality.

  59. Nice article, thank you. I’m guilty of using ‘luck’ and ‘random’ interchangeably above; mostly this was motivated by the writer’s desire not to keep using the same word over and over again, but there is a chance that it might confuse the reader. I was also DEEPLY shocked to find that Paul the Octopus’ World Cup predictions were just chance! 🙂

    I’d agree with you (but with little empirical evidence) that luck is likely to be more of a factor for the top and bottom end of the spectrum of X-Wing players. It kind of makes sense. Theorist’s scale of skill vs luck is useful too, I agree, if also arbitrary. [I would love to have the right methodology to accurately determine what that ratio was, as it might give some insight into human intelligence.] And you are right, vdude, that X-Wing holds a special fascination for all of us here, which prompts some excellent debates, where other games do not.

    There is something about it’s emergent properties and (in my opinion, though one probably shared) the simplicity of balance between chance and choice; David can take down Goliath. We see it every time a TIE Swarm flies.

    The article link you posted reminded me of the illusion of the “hit picker” common in the days of the “studio system” in Hollywood’s golden age. While I cannot find the reference to hand, researchers concluded that the so-called star producers, who could “smell a hit” were, upon statistical analysis, no better than the “stinkers” who got fired after their first flop. Sorry I cannot provide a reference.

  60. The premise of this article was because I disagree with your definition of Hindsight Bias. It’s not my definition, that’s the scientific community’s definition. Generally i disagree with the premise because you are guilty of the same transgressions in which you are accusing Sunny.

    The term you’re looking for is confirmation bias. It’s when an observer prioritizes information or interprets results in hopes of confirming a preconceived notion. That is Sunny felt X-Wing is not a game of luck and found Paul winning confirmed that. It would be hindsight bias is Sunny claimed he knew all along that Paul would win back to back Worlds, even though he had no way of knowing it would happen.

    Also your math for Paul’s chance of winning twice in a row is way off. Picking the winner for Worlds is an independent trial. Assuming there are exactly 256 players in each event and each event contains the exact same set of players that’s 1/256 *1/256 = 1/65536. It seems fairly unlikely someone will repeat.

    I think the only real argument to make here is as 2 players’ skills converge, random events, such as dice and damage cards, have a much greater impact on the outcome. Let’s assume both players are playing optimally. The result is neither player makes mistakes and dice random events are likely to decide the outcome. The problem is no one can play perfectly all the time and few players can identify optimal play. That leads many players to incorrectly blame luck on the outcome.

    All the same, thank you for the interesting read.

  61. I like what you have to say. I do think as time goes on, and thanks to the champs helping so much, that there will be more players that can decide the best line of play. It is like any other game/sport, it boils down to how much time you want to give it.

  62. Hi Sunny. Apologies for the delay in replying; I’m trying to go through comments in order of posting, while fitting it in between bouts of other work. My response to your original article was indeed because it appeared to paint in black and white the fact that Paul’s victory “proves without a doubt that skill overcomes dice in the long run” and “Heaver repeating at Worlds is the ultimate nail in that argument”. These are both pretty definitive statements, as is the concluding remark, “You’re not lucky; you’re good.” Hence, the need to take a step back from these claims.

    While I don’t mind the odd bit of rhetoric, to say that you “attribute no luck to winners” seemed perfectly justified, given the language of your article. I do concede, however, the fact that we will never know if you would have written a similar piece if Morgan or Keith Wilson had won instead; that’s the trap of hindsight bias. However, to dismiss this argument by ‘reductio ad absurdum’ comments about “Bob from my local game store” winning Worlds, doesn’t wash. We both know that there is too much skill involved for said ‘Bob’ to get past Flight One. That would only be a possibility if the game was just luck. I think that it likely that in the cases of any of the top four players winning, I would probably have been tempted to pen a piece extolling the triumph of skill over luck, if I had wanted to address the issue at hand; namely your concern that people dismiss their own responsibility when they lose. This is, after all, the main thrust of your article: encouraging new players to reflect upon their choice, rather than blame the dice.

    Your comment on Theorists Skill vs Luck Scale – we must come up with a good name for that – placing the line at 9:1 (you use 90% in your comment, but I tried to put it in Theorist’s language) is just as arbitrary as his 4:1 or 6:1. I think the only common ground that we have here is that we all agree it is N:1 where N>1. Statistically, if Paul had a 0.9 probability of winning a game, to win 11 in a row is ~1/3, meaning that he had about a 33% chance of winning Worlds. It might be worth considering revisiting the best players and seeing where they would be seeded, to see if Theorist’s scale could be used as a measurement or way to predict their success.

    At least you have asked whether I “feel no value in competing”, rather than just assuming that playing the game for fun implied I believed the game to be “so luck based” I wouldn’t bother to try to win. The answer here is “No”. Neither do I think the game is dominated by chance, nor do I ever deliberately play to lose. Playing for fun, rather than picking this week’s sure fire winning meta, is at the heart of “Fly Casual” surely? You do go on to make assumptions though, over my choice to not fly TIE Swarms, and my discussion of Heaver’s choice of Fat Han, (but with an R2D2 twist,, which I didn’t go into details over). This was an aside, and linked to the criticism (not mine) of one player who chose to ‘fortress’ up his list, because it was the best way to deal with a ‘Rock beats Paper’ opponent (that is one who’s list the fortress player’s ships would otherwise have been particularly vulnerable to). From reading Paul’s Worlds 2014 report, he made great efforts to choose a list that combated the most likely serious threats, and then spent a good deal of time researching and practising with that list. I’d say that was the ‘nicer’ side of WAAC, in that he put in every effort to maximise his chances. The fact that Paul took a unique take on Fat Han is to his credit, but it doesn’t belie the fact that Fat Han is considered the current ‘go to’ list; hence the World shortage of Falcons.

    Regarding the Radio Lab podcast, I’ve edited the OP, because the link provided isn’t the one that I was remembering. Thank you for pointing out the error. Sadly, despite trying, I haven’t found the correct link, although I suspect it was probably one of the Freakonomics podcasts. However, I have posted an article that discusses many of the same issues. Your breakdown of different types of poker is insightful, and I’d agree that there is a different skill/luck ratio to that of X-Wing. The difficultly is in being able to measure that. Poker is based upon playing cards, and there is a way to calculate the chances of a particular winning hand, but X-wing introduces many more choices in a single game, and the added variability of different lists, etc, makes the game harder to judge objectively.

    Finally, thanks for the clarification over the purpose of your article. While I don’t think you have shown that there is a wave of new players blaming dice, as you imply, the message of reflecting upon your performance, rather than blame the dice, is a great one. It just doesn’t need the debatable evidence of Paul’s second victory. I do think that Piqsid’s article does the issue better though. Sorry. Part of my problem with your article, and your response here is that you constantly appeal to apparently self-evident data – the extreme argument that the top four couldn’t have got there by luck, in this case – which presents a false dichotomy: Either luck is a negligible factor that skill overpowers, or it’s all just random chance. You place yourself at one end and, I believe, me at the other, when I think we are much closer to each other. I don’t need to tell you “with a straight face…that luck got [the top four] there” as this is blatantly ridiculous. I just need to say that luck is not absent either.

  63. Congratulations. Are you a medical doctor or a PhD? I got mine in 1999 from the University of Glamorgan, after eleven years (!) working on AI in games and simulations. Having the Dr. in my ID (and all my online IDs in fact) isn’t intended to be affectation, but is in part to distinguish myself from the other “Mike Reddy”s of the world – one a football player, one an artist, and one a UK government official – and partly historical. I worked for Portland TV on their video games channel, and bizarrely there were many “Mike”s in the studio. We had “Camera Mike” and “Commentator Mike”, so I became “Doctor Mike” and it stuck.

  64. Thank you for the clarification. You are right that “confirmation bias” also applicable; I think it depends on the order in which you see things. However, I suspect that we could get into an argument of semantics if we proceed. Let’s just agree that I might have mislabelled the bias, but that we really need to address whether the evidence supports the hypothesis.

    You are correct that if it was truly random, which I don’t think anyone is arguing, that the chance of winning twice is 1/65536. The 1/256 was in reference to a single tournament. However, I was trying to show that someone with a 60% chance of winning a game had the same chance of winning eleven games as just picking someone randomly from the two flights. This was, of course, an analogy, but intended to show that results do not always prove causation.
    [Edit: Of course, as a colleague reminded me, the probability of winning, given a win, in a truly random “lottery” style event, or the conditional probability, would still be 1/256 in the same way that tossing a Head on a coin is the same no matter how many heads have been tossed before. Yes, this still only applies to the situation where skill is absent. Conditional probabilities involving skill are the basis for seeding in tournaments.]

  65. I think again we are disagreeing about different things. I like the juggling analogy, but it would be better done with two jugglers sharing balls, given that X-Wing is a two player game. However, no claim of a person’s ability to juggle without any preparation is being made. I was just trying to make a case that statistical significance is missing. Others have quite rightly argued that as the player skill rises, the factor of luck also returns. This is also true down at the bottom end of player skill, although I’d like to think I wouldn’t blame the dice for my lesser ability. For Sunny to argue that Paul’s performance is an exemplar of hard work and minimising the odds, I would agree. For him to state this proves X-Wing is a game of skill, not luck, at all levels of play is doubtful.

  66. I think we can all agree that Luck is a factor in the game. The very fact that we roll dice adds that element. We can examine dice rolls and take a look at how they oscillate in values. Mathwingers love this. What we don’t have is a way to measure Skill. A tournament, even one as large as Worlds does not really give enough data to go on for a more precise view.

    We can look at Paul’s history. Worlds 2013 champion, Worlds 2014 champion, Top 16 at nationals for 2 years in a row( It could be top 8 for both but I’m not sure). Vassal Champion. Making the top cut in Galactic cup and the TCO.

    While this again is a small data set, I don’t think it can be over looked. Consistently making the top cut in major events tells us something.

    There are more area’s of X-Wing where Skill is the major factor.

    Squad building = Skill
    Strategy with Placement of Ships and Asteroids = Skill
    Understanding how to attack your opponents squad = Skill
    Flying/Maneuvering = Skill
    Attacking/Defending = Luck
    Action Selection/Use = Skill
    This list is by no means exact but just to demonstrate that more area’s of the game are determined by skill then luck.

    I feel that Paul’s repeat at worlds had more to do with Skill then Luck. The reason I think this is because of his history of success. I understand that it’s still a small dataset, in terms of number of games but it’s the consistency of success that adds weight for me. I’ve watched Paul play and he doesn’t make many mistakes. This part is very hard to quantify. Not sure how to express this point other then he makes the right plays when he needs to, and rarely gives up his advantage. The prep work that he did before the tournament lets him offset fatigue that might effect other players. When you know your opening and have it down, then you are less likely to make mistakes. This sets you up for later in the game. Prep work is one way to mitigate mistakes that you might make from fatigue. This is a skill.

  67. I don’t think that there is any doubt that the final placings, with Paul at the top had very little to do with luck and everything to do with skill.

    You use lack of statistical significance to support your lack of confidence in the importance of skill over luck. What you need is more data.

    There were over 250 people in that tournament, some of them were national champions, some of them regional champions, some of them had proven themselves on Vassel tournaments, etc, and many of them were just local enough that they could turn up and make up the numbers. What more data do you need?

    I think it is fair to say that probably more than half the field (maybe a lot more) would have been American (local) players, with no real credentials, in a horse race they would be considered outsiders, you have no reason to suppose they would do well. If this was a game of luck you would expect 2 of the final spots to go to these people. Even if luck played a minor role you might expect one.

    So how many local outsiders made the top 4? Paul Heaver, standing world champion, Morgan Reid and Keith Wilson, neither local, both National champions. I know nothing about Ira, is he a local with no credentials that relied on luck, or is he an established player with a proven reputation? Someone else can probably answer that. Interested to know similar about the other Top 8/16/32 players.

    I think if someone went through the list of entrants, ranked them according to past acheivements, and tried to give them odds, like a horse race, we would see the outsiders falling by the way-side very quickly, and the top 32 would be predominantly filled with named players and only a few outsiders. That may help to give significant insights into how much luck is required to get through.

  68. Thanks. sorry for not responding earlier; I assumed the thread had petered out; it was a little tongue in cheek, being mostly a commentary on not assuming too much from limited data. However, what could be useful is an estimation of the probability of winning – effectively a measure of how good a player is – from past performance. You cannot completely exclude luck, as it is a stochastic as well as hidden knowledge game like poker. Even the best players cannot guarantee to win, and at the elimination stage, it only takes one loss to have you out of the tournament.
    The discussion here, and in the original post, also led me towards thinking how effective a sort algorithm Swiss is, assuming that there was some hypothetical order of ability that the approach is effectively sorting. Here are some relevant references:

  69. I believe the issue was your assumptions. MoV is likely mucking this up, since all the people win have the same MoV and so do the people who lose every round. So it is impossible for it to separate these people aside from eventually directly pairing them with each other.

    If you put in arbitrary, but not 100-0, values each time, I’d bet it would separate/sort much more quickly.

    If X-Wing still used SoS, this would likely have converged much more quickly. If you wanted to give it another go, I’d just put that # 1 always wins 100-0, #2 wins 99-1 (unless playing someone < him), etc. This will allow the tie-breaker to actually help separate the field as it is intended to do.

    A another idea would be to run the normal number of swiss rounds for 64 people (6) and determine how many 1 loss/2 loss/etc people you should have and see if they are in that range of people at the end of 6 rounds. So you might have 1,3,2,4,and on but they would be in the correct range bands for their expected W/L ratios. You should note though that depending on who they are paired with when is also a factor. If #2 is paired against #1 and #4 is paired against #3 in round 1, and then #2 and #4 are paired against each other round 2, #4 will have 2 losses, despite being a player that would otherwise "should" have been in the top 4 with only a single loss.

  70. I did what I suggested for 16 people (#1 wins 100-0, #2 wins 99-1, and so on) and at the end of the recommended amount of rounds I got the Top 4 to be 1,2,3,5 with 4 being just below #5 on MoV (but ahead on SoS). by only 2 points. I expect if I increased the MoV values for each round to be more realistic, #4 would have made it, regardless of pairing order.

    In fact the entire field was very close to being in perfect order after 5 rounds.

    So in short, my simple conclusion is the system and application work fine as-is.

  71. You had me interested in what you did here so I just did 6 rounds using your criteria. After the 6 rounds the top 16, in ranking order were:

    Name Score MoV SoS Record Byes

    1 30 1200 110 6 / 0 / 0 0
    2 25 1000 130 5 / 1 / 0 0
    3 25 1000 130 5 / 1 / 0 0
    5 25 1000 110 5 / 1 / 0 0
    7 25 1000 100 5 / 1 / 0 0
    10 25 1000 95 5 / 1 / 0 0
    16 25 1000 85 5 / 1 / 0 0
    13 20 800 125 4 / 2 / 0 0
    4 20 800 115 4 / 2 / 0 0
    8 20 800 110 4 / 2 / 0 0
    6 20 800 110 4 / 2 / 0 0
    11 20 800 110 4 / 2 / 0 0
    9 20 800 110 4 / 2 / 0 0
    24 20 800 100 4 / 2 / 0 0
    18 20 800 100 4 / 2 / 0 0
    14 20 800 100 4 / 2 / 0 0

    #12 (in 17th place) and #15 (in 28th place) missed the cut with records of:

    12 20 800 90 4 / 2 / 0 0
    15 15 600 100 3 / 3 / 0 0

    Who did they play each round (1/2/3/4/5/6)?

    #12: #62 / #46 / #15 / #9 / #10 / #21

    #15: #60 / #21 / #12 / #10 / #19 / #14

    #24 and #18 being the outliers here in the top 16 I looked up who they were paired with in each round (1/2/3/4/5/6):

    #24: #21 / #62 / #42 / #41 / #31 /#16

    #18: #35 / #53 / #6 / #20 / #23 / #10

    So #12 lost out because of strength of schedule and #15 because of the 3-3 record. #24 was given favorable match-ups in rounds 2 to 5 giving him the 4 wins needed to make top 16 with interestingly a higher strength of schedule than #12.
    Looking into this I found his opponents also did well #31 placed 19th and #42 placed 20th increasing his SoS, whereas #12’s opponents #46 placed 47th and #62 placed 61st, which wasn’t helping #12’s SoS at all.

    So in this way it looks as though being a player in the 10 to 16 range while getting a 1st round match against a bottom feeder (#12 v #62 and #15 v #60) really hurt their chances of making it in.

    Thus I do not think your findings after 6 rounds were odd. The random pairings in the first round (and second round) seems to be the reason why we can see outliers making it into the top 16 and others we’d expect to be in getting put out.

    As for continuing the investigation as you did, I did not. I can see why 12+ rounds would start to break the system, but a tournament of 64 players would never be that many rounds so I think its a non-issue.

  72. It would be worth trying this but using a more intelligent method for MoV. Something like:

    win margin = min(100, 4*((lower player #) – (higher player #)))

    Of course this results in partial wins for anyone ranked within 2 of each other, so if you want to enforce full wins then instead try:

    win margin = max(12,max(100, 4*((lower player #) – (higher player #))))

    No need to use cryodex for this, this would be better suited to writing an actual script fr it. Then you could easily iterate many times. You could even add a uniformly distributed random number into the win margin.

    win margin = min(100, 4*((lower player #) – (higher player #)) + random([-25 to +25]))

    I would whip this up myself in MATLAB, but… dissertation writing calls. You could play around with the constants and see what you get for results.

  73. It seems odd that modifying the MoV is necessary to ensure that the Swiss produces an eventual rough ordering. I can see from evidence above (which I am not questioning) that this is needed, but in the real world, 100-0 etc, are more likely (especially in Vassal, where there is often no time limit), so the question remains. How good is Swiss as a sorting algorithm. We can certainly ‘fix’ it in this hypothetical case, but I suspect that it is still only good at giving us a best guess, in comparison to a full league where everyone played everyone else. Has anyone else reproduced the bug of a single player having to play two games at once?

  74. Games going to time are very frequent in tabletop tournaments, especially with the rise in popularity of the game, especially when 16+ players show up and the need to reach at least 4 rounds of swiss with at least a top4 or top8 elimination rounds. Those conditions all push towards short 60 minute rounds and make 100-0 less common.

    Vassal is the odd beast in this situation, because you usually run tournaments across weeks/months and 4 hour games are possible and the higher number of swiss rounds help the algorithm have enough “time” to work its thing.

  75. Hey Doc, could you send me your save file from the test? I’d be interested in seeing what went screwy in round 13 that got one player in two different matches. If there are problems in my software, I want to fix them. I can’t do that unless people report these issues to me.