This makes it even more important to have all of the best tools at your disposal during the development process! With zero playtesting before release, I was to accurately predict ship performance better than the developers in the above cited instances (I am judging that I did a better job in these instances than the developers because I am assuming that they intended for the ships to be costed / balanced appropriately). I'm obviously not saying to throw playtesting out the window, on the contrary. But rather you really want a more well-rounded set of tools to analyze balance than simply throwing ships on the table and getting a relatively small sample size of empirical performance.

I do not want to be rude or anything else, but I have the impression that you feel the need that you are good at stats/that your model has predicted things very often. You do not need to state it at every post, despite the obvious interest that you work has (and that I recognize). Thanks.

My apologies. The point was to use it as a premise to reach the conclusion that:

Part of the reason we are in such a strong paper / rock / scissors dynamic is because the developers have relied on empirical play testing data (which is inherently a small sample size) without fully understanding the analytical mathematical theory.

And by the way, the HLC Outrider is going to be the new "lizard" or "Spock", and it will be worse than the ACD Phantom or C-3P0 Falcons. The only real counters to it will be higher PS maneuvering, and possibly swarms.

It will be interesting to see and it certainly seems like Dash is going to be very tough. However, beyond whatever the numbers say on this one, there are ships that the Outrider just plain will not be able to deal with in the endgame because of that donut, and it's more ships than are "easy" counters to the phantom, increasing in number as the PS of the Outrider drops. Dash makes that a lot more difficult with his ability to move freely through obstacles, though.

Beyond that, even if it is the new Lizard/Spock, it may have a small amount of time to shine based on the "anti-turret" card and whatever it does (which is admittedly a list building strategy), and that small amount of time may not include a premier event if it's not released by World's.

My intention going forward is to drastically reduce my involvement with the forum, hoping to be more productive on my dissertation as well as less frustrated overall, but I want to address a few points.

MathWing has been correct in accurately predicting how well almost all of the ships will do in the competitive tournaments, before the ships have even been released , based on objective numerical analysis . For example:

If you're saying that FFG accomplished perfect points pricing on their ships and upgrades, then you need to say that more clearly as well as back it up with some math. From everything that I've seen, I think FFG did pretty well, but by no means perfect as MJ has demonstrated.

Yes, I really wish that they had strong analytical math skills. While it would not change the paper / rock / scissors dynamic, there would at least be relative balance between all of the ships. In each of waves 1-4, there has been at least 1 strong "dud" ship that was entirely preventable.

• wave 1: TIE Advanced. overcosted by 4 - 4.5 points relative to a TIE fighter, and lacks a useful role.
• wave 1: generic X-wing: overcosted by 1 point, relatively minor effect in waves 1-2, but now obsolete by B-wings and Z-95s.
• wave 1: Y-wings: overcosted by 2 points if they don't equip a turret. A better implementation would have been to drop Y-wing costs by 2 points across the board and then make a Y-wing only turret slot upgrade that costs 2 points, or equivalently a -2 cost turret "refit" that eats the turret slot.
• wave 2: A-wings: overcosted by 2 points before Refit in Rebel Aces.
• Outer Rim Smuggler: clearly overcosted by several points based on its jousting value relative to the named (different stat line) YT-1300, and inherently lower upgrade utility on a cheaper ship.
• wave 2: TIE Interceptor: low PS generics pay too much for boost action, since they can't really take advantage of it.
• wave 3: Bombers: the ship itself isn't terrible, but it needs cost effective ordnance to be useful, which doesn't exist. This problem stems to wave 1.
• wave 3: Rebel Operative: definitely pays a huge price for its crew slot compared to a generic Y-wing.
• wave 4: generic E-wings: overcosted by 3 points.
• wave 4: TIE Defender: generics are slightly overcosted, although admittedly the white K-turn is hard to place a value on.

I say with confidence that if FFG doesn't have the kind of analytical skills you mean on their design staff, they at least have access to them in the playtesting phase. There are a lot of people with strong quantitative backgrounds in X-wing--like lots of techy, geeky games, its fanbase seems to have a disproportionate number of the terminally well-educated--and it's very improbable that no one is giving them feedback from a MathWing standpoint.

As for your list, I think it's possible to synthesize problems in X-wing ship prices to the following:

(1) Having 2 Attack sharply limits the increase in value from improving other aspects of the ship. A Prototype Pilot is clearly not worth nearly half again the price of an Academy Pilot despite having a better action bar, a better dial, and at least 25% better durability.

(2) Increasing PS by 1 is not always worth 1 point in cost, but it's often priced that way. Even a looser version of this stricture, where the appropriate cost for a valuable unique pilot sets an approximate lower bound for the cheapest generic pilot, can cause problems.

(3) Ordnance is unreliable and expensive, which means it will likely be unpopular even if it's effective on average (and sometimes it isn't even that). Ships that are or were clearly priced with ordnance in mind, like the Prototype Pilot + Concussion Missiles that can just barely not be fielded in fives, are likely overpriced as a result.

Those three hypothetical roots explain all the issues you noted, plus at least a couple more you didn't address... particularly if you also account for the fact that, as we've discussed extensively elsewhere, I have some fundamental issues with your modeling approach.

With zero playtesting before release, I was to accurately predict ship performance better than the developers in the above cited instances (I am judging that I did a better job in these instances than the developers because I am assuming that they intended for the ships to be costed / balanced appropriately). I'm obviously not saying to throw playtesting out the window, on the contrary. But rather you really want a more well-rounded set of tools to analyze balance than simply throwing ships on the table and getting a relatively small sample size of empirical performance.

I have an issue with this because--and for the second time in this thread, I'm trying to be very careful with my language to avoid implying something I don't mean--you are placing a great deal of faith in mathematical modeling without acknowledging how much of all modeling is an idiosyncratic "art", and how much of your jousting model in particular is based on assumptions that very few people pay attention to or attempt to verify when they're discussing value.

We disagreed fairly vehemently over the Defender, for instance, and I think the stance you appear to be taking here--that not only do you think it's overpriced, but that it's objectively overpriced--is potentially in error.

Your conclusion about the Defender is that under a particular set of assumptions about the action economy and the ranges at which attacks are made, and given a particular relationship between a ship's general cost and its value, the Defender appears to be overpriced. I disagree strongly with at least one of those assumptions and don't feel comfortable relying on several of the others; obviously you believe those assumptions are robust. But there's a difference between "I feel confident about my assumptions as representative of a general use case" and "my assumptions are objectively true", and I want to flag that difference for your review.

And, finally, back to X-wing rather than the epistemology of quantitative modeling:

And by the way, the HLC Outrider is going to be the new "lizard" or "Spock", and it will be worse than the ACD Phantom or C-3P0 Falcons. The only real counters to it will be higher PS maneuvering, and possibly swarms.

I'll say it again: I think rock-paper-scissors is a fundamentally inappropriate comparison for a game that features a nearly limitless number of combinations for each player's list. If I can take a list that's part rock and part scissors, as well as a list that's neither rock nor scissors, then RPS isn't a particularly useful way to look at the game.

To the degree that it is true of the current metagame, I think Wave 5 will relieve the problem (at least if the problem is defined as over-representation of two list archetypes in the Falcon and Phantom), and Wave 6 is going to shatter whatever crystallization has taken place.

And here we go again with the armchair game designing. It's not to say that FFG doesn't make mistakes, but your analysis is a small vacuum of what's currently releasing now or in the near future, and at the 100 pt level.

A release cycle probably takes in the neighborhood of 1-2 years and so there are a number of things of future waves and versions of the game that aren't really known to us that are also taken into account. When wave 1 was being set into stone, I'm sure wave 2 was actively being developed, along with concepts and directions for wave 3 and epic play.

I believe you have made this point before. There are 2 major flaws with this reasoning:

1. It assumes that the developers would intentionally release a ship knowing that it would be useless in the competitive meta .
2. We are 4 (almost 5) waves into the game, and you have not cited any empirical evidence of how under performing / overcosted ships were suddenly redeemed by later upgrades. I challenge you to put your theory to the test and see if ANY of the below cited examples have been "fixed" by introduction of later ships / upgrades, and that this was intentional from the outset. ^*

* Chardaan Refit doesn't count, since it is a blatant point cost reduction, which the A-wing should have had in the first place, as predicted by the analytical theory.

In each of waves 1-4, there has been at least 1 strong "dud" ship that was entirely preventable.

• wave 1: TIE Advanced. overcosted by 4 - 4.5 points relative to a TIE fighter, and lacks a useful role.
• wave 1: generic X-wing: overcosted by 1 point, relatively minor effect in waves 1-2, but now obsolete by B-wings and Z-95s.
• wave 1: Y-wings: overcosted by 2 points if they don't equip a turret. A better implementation would have been to drop Y-wing costs by 2 points across the board and then make a Y-wing only turret slot upgrade that costs 2 points, or equivalently a -2 cost turret "refit" that eats the turret slot.
• wave 2: A-wings: overcosted by 2 points before Refit in Rebel Aces.
• Outer Rim Smuggler: clearly overcosted by several points based on its jousting value relative to the named (different stat line) YT-1300, and inherently lower upgrade utility on a cheaper ship.
• wave 2: TIE Interceptor: low PS generics pay too much for boost action, since they can't really take advantage of it.
• wave 3: Bombers: the ship itself isn't terrible, but it needs cost effective ordnance to be useful, which doesn't exist. This problem stems to wave 1.
• wave 3: Rebel Operative: definitely pays a huge price for its crew slot compared to a generic Y-wing.
• wave 4: generic E-wings: overcosted by 3 points.
• wave 4: TIE Defender: generics are slightly overcosted, although admittedly the white K-turn is hard to place a value on.

You have labeled the analytical approach as mere "armchair game designing", while altogether avoiding discussing anything particular about it. If you can't cite a single example of your theory working in action, then one might as well label your comments "armchair forum commenting" by the same standard. I would rather have a discussion on the particulars. Will Y-wings without turrets be viable after wave 6? Will the TIE Bomber be relevant with new ordnance? Will the TIE Advanced have a unique and cost effective role eventually? Will we ever use generic X-wings again, in the game called "X-Wing Miniatures"? These are all interesting discussion points and constructive discussion can be had in regards to how the mathematical theory can, or cannot predict the effectiveness of such changes.

The game has never been exclusively about jousting, and the mathematical theory has always gone well beyond merely calculating the jousting values. That said, I agree that it is very healthy for the game to add new dynamics, if all they did was add different permutations of stat lines, cost, and dials, the game would become too homogeneous.

And here we go again with the armchair game designing. It's not to say that FFG doesn't make mistakes, but your analysis is a small vacuum of what's currently releasing now or in the near future, and at the 100 pt level.

A release cycle probably takes in the neighborhood of 1-2 years and so there are a number of things of future waves and versions of the game that aren't really known to us that are also taken into account. When wave 1 was being set into stone, I'm sure wave 2 was actively being developed, along with concepts and directions for wave 3 and epic play.

I believe you have made this point before. There are 2 major flaws with this reasoning:

1. It assumes that the developers would intentionally release a ship knowing that it would be useless in the competitive meta .
2. We are 4 (almost 5) waves into the game, and you have not cited any empirical evidence of how under performing / overcosted ships were suddenly redeemed by later upgrades. I challenge you to put your theory to the test and see if ANY of the below cited examples have been "fixed" by introduction of later ships / upgrades, and that this was intentional from the outset. ^*

* Chardaan Refit doesn't count, since it is a blatant point cost reduction, which the A-wing should have had in the first place, as predicted by the analytical theory.

Yeah, in counter to the same points you're making here.

1) Those are your assumptions. I don't assume them to be useless or that they design completely to the 100 pt. metagame. I assume they do overcost certain ships/slots for the future for some design headroom.

2) I'm not the one claiming to know how FFG designs games and that they don't do mathematical analysis. That ball is in your court. You're the one making assumptions. I have given scenarios previously of what could have been the reason why the Tie advanced and A-wing were overcosted. However, another example is ordnance/bombers. Purported as being overcosted in 100pt. games, developed before epic play, and actually well costed in epic play.

I have discussed a lot of details of game design and release cycles. You're the one that seems to think it should only involve math without ever having designed a game of this scale and expansions, so yes you are "armchair game designing" here.

Really, can we see this mathematical analysis? What's the mathematical value of a phantom decloaking?

I've got two questions:

1. What was the truly creative list? (Pure curiosity)

2. What are your criteria for a list to be "creative"? (I ask because, in my humble opinion, adding Jan to a Falcon list is a creative twist on a well-known list archetype, does that make it creative enough?)

Although there's in principle hundreds of thousands of list combinations, there's a small number of list archetypes.

Are you (or is anyone) invested enough in the question of whether games are won or lost for the most part before the ships hit the table to get behind an empirical test that would help illuminate the answer?

I think the Blue/Biggs/Roark/Tarn list was pretty original. It's a weird list that I'm not sure is actually a good list, but he obviously managed to make it work.

"Creative" to me at the very least doesn't start with a core of the obvious best ships in the game right now. Running Fat Han with a Chardaan'ed Prototype instead of a Bandit isn't really creative. It's not a boolean - things can be more or less creative. Fat Han is basically the same as it's been run since Han Shoots First - a nasty Falcon with as much support as you can cram into it. Phantom+Mini-Swarm really isn't very creative, it's the same as has been done with Firesprays for a long while. It's certainly possible for there to be bursts of creativity in lists, but overall it's bleh.

Edit: And just to make sure nobody takes this out of context, there's nothing wrong with running an "uncreative" list. You aren't a bad person for flying a Phantom+Mini-Swarm or Fat Han with a Rookie and Bandit. The point I was making is that I take issue with the people who go on and on about how there's not anything so good about the Phantom or Falcons, that it's all in our heads, that there are plenty of other choices if we'd just stop being netdecking sheeple... and then turn around and take exactly those ships into competitive events.

As for the last test, I think it's a false question, for two reasons. First is that it's almost certainly an impossible task. But more importantly, why is it directed at the people who think build has become a major factor? We have a difference of opinion on it, and that's fine - are YOU invested enough in the idea that build has no influence to come up with and carry out an equally empirical test?

It is safe to assume that they are using some level of MathWing perspective, but the evidence is nearly irrefutable that their MathWing approach is not nearly as rigorous or accurate as my approach. Don't shoot the messenger! I'm just another (as you put it) "terminally educated" fan...

1) At a broad level, I could interpret that as the entirety of my analysis past jousting values. Since I model non-jousting factors (like the dial or System Upgrade for example) as linearly increasing overall combat effectiveness, it does inherently place more point value on, say, putting crew on a named YT-1300 than one a HWK-290. I have a special case for Boost where I value it depending on where the ship falls on the glass cannon vs tank scale. (It being worth more on a tank, obviously)

2) Agreed, but the examples cited above are all the lowest PS versions of the ship, so the effect between PS1 and PS2 is negligible.

3) Agreed on ordnance being too costly for what it does. It has essentially killed the TIE Bomber, whose only fault is that the Imperial Munitions Factory has a monopoly, and never has a sale on missiles or torpedoes. So sad.

Your conclusion about the Defender is that under a particular set of assumptions about the action economy and the ranges at which attacks are made, and given a particular relationship between a ship's general cost and its value, the Defender appears to be overpriced. I disagree strongly with at least one of those assumptions and don't feel comfortable relying on several of the others; obviously you believe those assumptions are robust. But there's a difference between "I feel confident about my assumptions as representative of a general use case" and "my assumptions are objectively true", and I want to flag that difference for your review.

Thanks for the feedback. There is definitely room for improvement on modeling different action economies and range bins, on both the mean and standard deviation. I already take a multiple-vector approach on the kinds of ships that are present in the meta, and it would be good to extend this to the action economy and range bins as well. So the efficiency predictions already have a +/- range associated with them, but there is room for improvement. In this regard, you can substitute " Defender " above with " any ship ". The Defender is a tricky bird to value specifically because of its unique white K-turn, but the above points of variance of action economy and ranges apply to all of the ships. As I'm writing this, I'm thinking that having at least 3 variables would be good: ship types, ranges, and action economies. I'll calculate the value using the central value for each variable, the worst case value, the best case value, and the mean value. If I take 5 uniformly distributed data points for each variable, that's 125 samples per ship, which should look like something approaching a bell curve.

But this doesn't change the validity of the underlying methodology, only the error bars on the efficiency estimates.

1) At a broad level, I could interpret that as the entirety of my analysis past jousting values. Since I model non-jousting factors (like the dial or System Upgrade for example) as linearly increasing overall combat effectiveness, it does inherently place more point value on, say, putting crew on a named YT-1300 than one a HWK-290. I have a special case for Boost where I value it depending on where the ship falls on the glass cannon vs tank scale. (It being worth more on a tank, obviously)

So you're saying you take non-numerical and situational abilities like, dials, slots, boost, etc. and assign them arbitrary values (without play-testing to see their actual value) and call it sound mathematical analysis?

So the efficiency predictions already have a +/- range associated with them, but there is room for improvement. In this regard, you can substitute " Defender " above with " any ship ". The Defender is a tricky bird to value specifically because of its unique white K-turn, but the above points of variance of action economy and ranges apply to all of the ships.

I'm picking on the Defender because I think it's the most prominent example of mis-fit in the model being interpreted as weakness in the ship itself--but yeah, the issue of how well your assumptions matchthe reality of a particular match between two players, each with a single list, definitely hangs over all of your ships.

But this doesn't change the validity of the underlying methodology, only the error bars on the efficiency estimates.

I don't want to get in the weeds on this (because it's rather arcane, because it's diverging from the topic, and because I'm busy this afternoon), but I think the largest vulnerabilities remain these:

(1) The exponential model is as reasonable as any and more than some, but the choice of exponent is essentially arbitrary. In combination with the fact that in general model fit decreases as prices increase, that suggests to me that either the exponential model is inappropriate or the exponent is wrong.

(2) Many of the adjustments--for dial, action bar, upgrade bar, etc.--largely amount to individual adjustments "artificially" altering fit for a subset of data points. This treads very close to post-hoc adjustments for individual data points, which is a bad idea for obvious reasons.

" Those are your assumptions ". Your pronoun reference is ambiguous. What assumptions are you referring to? My point 1) is merely pointing out the logical necessity of your theory. It says nothing about what I believe.

" I don't assume them to be useless ". Well... all of the above examples are very very poorly represented in the competitive meta. We have quantifiable data for that, so we don't need to make any assumptions one way or the other.

" I assume they do overcost certain ships ". I equate "overcost" with "useless in the competitive meta", so YAY! We agree!

1. You are putting words in my mouth. This is ill-mannered, and also a very poor debate tactic, as it is trivial to now prove you wrong.
2. I have repeatedly and specifically stated otherwise, including this statement only a few posts above yours:

You either missed this point, or chose to ignore it. Either way, it is not helping your case...

I don't have anyone on ignore on this forum, but I'm not going to waste my time debating with people who not only ignore what I say, but then put the exact opposite words in my mouth. So, keep it up and you will likely be the first on the list, and you will stop getting any responses from me.

Really, can we see this mathematical analysis? What's the mathematical value of a phantom decloaking?

http://community.fantasyflightgames.com/index.php?/topic/100360-using-lanchesters-square-law-to-predict-ships-jousting-values-and-fair-point-values-work-in-progress/

Really , please keep it civil.

Non-dice factors (stat line) such as unique dial maneuvers (like white K-turrns or red 0) and unique actions (cloak) are more difficult to place a relative value on, since there is no comparative baseline with which to balance between ships. Other actions such as Target Lock are easier to quantify since you can numerically compute the increase in expected damage output.

Therefore the certainty of predicting the overall efficiency of some ships is lower than others, and again I have been up front about this. However, this does not mean that the entire method is worthless, especially for those ships that do have common functionality. You seem to imply that because it is difficult to mathematically quantify one specific action on one specific ship, that it renders the entirety of mathematics out the window.

In the case of the TIE Phantom, the pertinent point is that as long as the cloak action does not have a negative value (which it obviously does not), an ACD Phantom with higher PS than its opponent is extremely cost efficient, even before considering the benefit of a "free" 2 speed decloak barrel roll / forward. So in this case, the jousting value puts a minimum floor on the performance of an ACD Phantom (again, assuming ACD triggers before it gets shot at), and that minimum floor is already nearly the highest efficiency in the game.

I don't want to get in the weeds on this (because it's rather arcane, because it's diverging from the topic, and because I'm busy this afternoon), but I think the largest vulnerabilities remain these:

(1) The exponential model is as reasonable as any and more than some, but the choice of exponent is essentially arbitrary. In combination with the fact that in general model fit decreases as prices increase, that suggests to me that either the exponential model is inappropriate or the exponent is wrong.

Except for the exponent being arbitrary, agreed! The curve fit has been fixed, but unpublished on the old thread. (It is floating around on one of my posts here, somewhere.) High value ships now get pushed further up the curve now as they should, even for ships that are worth 50 points at PS1. Unfortunately for the expensive ships, I revamped the durability calculations, and they took an even more massive hit. Random trivia, did you know that the expected number of shots required to kill 1 TIE Defender is actually slightly less than the number of shots to kill 2 TIE Fighters, even though the Defender has 3 shields to absorb crits? That was an interesting result, but not entirely unexpected. Lots of cheap ships requires more kill shots, which means more potentially wasted damage from the attacker.

It's not completely post-hoc though, because changing the coefficient will change the value for all of the ships with that action. So for ships that have shared functionality with all the other ships (like the E-wing), its impossible to massage its numbers without drastically affecting several other ships at the same time.

Major juggler,

I'd like to point out a fundamental shift in gameplay that happened during wave 1. Multiple dogfights --> squadron fights.

The tie advanced was not (as drastically) overcosted in the initial days of wave 1. It became overcosted somewhere between nationals and worlds. (And vader was in the initial winning lists..) Initially the tie swarm was not flown as a block, nor with howlrunner, people out 8 ties on the field and had at it. Often, games were played with a left / right approach, and not clumping. ( look back to the earliest play reports and videos)

In the one v one scenario, the tie advanced did alright, as it's defensive efficiency was good enough to survive a dogfight and then move to the other side of the field. As the shift occurred, and overlapping arcs, and tight formations became the way to play, jousting with squads rose. Now, the defensive efficiency counts for less, and we start seeing it as overpriced.

The "value" of a ship is fluid, and for the play environment prior to this switch, I don't think it was so badly costed.

Ravncat, couldn't had been because people was learning the game at the time, to understand and learning to exploit the game mechanichs more efficiently? I wasn't there at the time, but "multiple dogfights" sounds to me as not focus firing, which is inherently a bad game plan (it was already for me when i started to play, and nobody needed to tell me about removing ships from the enemy to gain advantadges).

As game progresses, the game mechanics are more figured out, and people learn much faster how to be efficient with the new toys, and how to throw to the trash bin what is not.

Just going to throw out there that MajorJuggler's numbers (and forum opinion in general) are held in such high regard that it's easily possible that just by making a prediction the behaviour of the playerbase becomes biased to follow it. Exhibit A, TIE defender. It's making a comeback now people are playing with it but the initial defender was slated so heavily it was pretty much non-existent. Not saying that if Juggler said it was awesome there would be a surge in TIE advanceds, but where things are more finally balanced a good or bad perceived jousting value can make or break it for a ship's popularity.

Dread, oh, I believe that's exactly what it is. But that also seems like how the game was designed, and intended to be played. Of course the game changes with players and time, but those early balancing decisions were noted to have been preventable above. I'm not sure they were given the game state at the time. Imagine it had been conceived of as a jousting squad game by everyone right away. I think we would've seen different costs, but that's not something can prove.

Going into the Malaysian Nationals I too was looking at a field allegedly dominated by Phantoms and Falcons (I moved back to KL from Australia) and I spent weeks in the lead up trying to find an answer to it. The answer I came up with was Jan Ors in a HWK and I agonized over which ships I could use to support her. And it worked.

I almost didnt run Jan because it flew in the face of conventional logic and I actually had a more conventional list in my back pocket. Even though I had tested playing both with and against Jan and had other good players critiquing my list. But that is the nature of competition.

You fail to mention what the rest of your list was here. What was Jan supporting?

I'll save the suspense: A Falcon.

6 of the Top 8 lists included either Phantoms or Falcons, including yours. There was one swarm, and one list that was truly creative. One.

I'm not sure if it's rank hypocrisy or people just don't understand what some of us are trying to say, but this is getting old. I was checking over TC and saw an announcement for Hujoe's game starting, and decided to take a look at his lists. You know what Mr. "There's nothing wrong with the game you just have to fly what you want Phantoms and Freighters aren't all that good!" is flying for his two lists? One with a Phantom, and one with a -2400.

The apologists love to say that it's all about the players, that there's nothing wrong with the balance, that it's all just perception and mindless netdeckers. But when the chips are down, they go to the same ships that everyone else is concerned with.

the issue of how well your assumptions match the reality of a particular match between two players, each with a single list, definitely hangs over all of your ships.

How so? I have always said that the numbers are " meta-wide averages, and can and will differ in specific matchups ". Eventually I will extend the model to look at specific matchups. It's trivial to calculate the jousting values for specific matchups, you just use a different weighting for ships you're fighting against, to limit yourself to the ships on the field at that time.

I mean that, e.g., your assumptions about the action economy are essentially certain to be violated in a non-negligible number of matches. Upgrades violate these assumptions all the time, and differences in playstyle can change things as well. Likewise, the chance that the set of ranges for a particular game will correspond to your assumption is certainly small.

I'm not saying that you've done a bad job, but that:

(1) You're attempting to make a reasonable assumption about the metagame as a whole, but that's dependent on your version of the metagame, and of your sketch of that version. To put it in more technical terms, you have a number of model parameters that are highly sample-dependent.

(2) Even if I spot you the assumption that your assumptions about the metagame are an accurate picture, the vast majority of individual samples don't correspond to any measure of central tendency.

(3) It's obvious, at least to me, that there's an interaction effect between a ship and a player. I can't comfortably run a Lambda without an Engine Upgrade, for instance, and I've tried. Equally obviously, that's unmodelable, but it means that my record with Delta Squadron Pilots or HWKs or the Outer Rim Smuggler might be substantially better than your model indicates even if the model were spot-on in every respect.

2) I'm not the one claiming to know how FFG designs games and that they don't do mathematical analysis. That ball is in your court. You're the one making assumptions. I have given scenarios previously of what could have been the reason why the Tie advanced and A-wing were overcosted. However, another example is ordnance/bombers. Purported as being overcosted in 100pt. games, developed before epic play, and actually well costed in epic play.

It is not immediately obvious to me that Bombers are inherently more cost effective than a pack of TIE Fighters or Lambda shuttles for the same cost. It is an interesting, and probably the best (and possibly only) example of how a ship does better in epic play vs 100 point games. Although to be fair, we will have to wait for Imperials to get their own huge ships before we have a "balanced" play field, otherwise we run the risk of developing a baseline against a CR-90 that may well be overcosted in its own right.

So if we assume that ordnance is more appropriately costed in Epic and that makes bombers more appropriately costed, we can say that other ordnance carriers also gain a bit more of an edge and a bit more point efficient?

I have no idea how FFG internally does design, but you can make the logical argument that:

If A then not B.

B is true.

Therefore A is not true.

Again that doesn't take into account, planning for future releases, variations on the game, etc. which you haven't addressed. This assumes that the following are true:

C) FFG only designs for the 100 pt. meta.

D) They are only concerned with how a ship balances with present ships and abilities.

It's not exact opposite, but that's minutia. You're right, that was jab and I'm sorry for that, but your arguments have been based around the mathematics of the system or the mathematical evaluations of empirical data so it's hard for me to see what other empirical things you think a game designer to take into account. So I feel that's where you are thinking about game design from.

You're welcome to put me on ignore, I don't care. I'm more concerned with what newer players think of the things that I consider you are off-base on and I feel it needs a counter-voice at times. I feel that your data collection and mathematics are cool, but you interpretations of those statistics at times flawed or incomplete, but yet spoken as definite and factual.

I'm serious about this and it give me concern for some of your evaluations. If you are making value judgments about abilities that don't have a direct numerical representation without playtesting those abilities, in some ways I consider it worse than game designing with only playtesting and no math. Some abilities are no big deal, but I bring up decloaking because it has a huge impact on how effective a phantom performs, but can't begin to see how to numerically represent that without playtesting it.

I just can't keep up with this thread!

@Snova

85/15 match up? I would say that most mixed 4-5 rebel swarm builds (no falcons!), with dice and player skill being equal START at about that. Unwinnable? No, not by a long shot but I have to expect to significantly outfly my opponent or get lucky to have anything close to 70/30 chance over time.

Similarly, most 4-5 ship imperial lists have an 85/15 match up vs. Fat Falcon . Imperials are really feeling hamstrung by what is starting to amount to 2 ships less to build with (advanced and bombers are really just not happening)... 6 named tie swarm is the exception and it's 85/15 is probably a well flown phantom (probably 80/20).

The example you posted about taking a z swarm, a ship with a disadvantage to start with required out-flying an opponent for a win. You later go on to say "expecting to out-fly your opponent every game is a good way to lose".

I don't agree that my only choice is to quit the game. I know better. I can wait and see. My playing is much less right now due to schedule and location. My presence here has slowed down because I am in waiting mode. I do need to wait and see if the game state changes before I decide to fully buy in with a new faction too. You don't rage quit a game this good, but you do point out when you see a massive shift in it occur.

I suspect that the development team wanted to shake things up, I see the way they shook it up as bad for the game right now. Will it always hold? No, but with what info I have it looks like it is going to hold for at least another wave and with the delays we're seeing that could last well into this time next year before a change is seen.

I don't like it, I'm not delusional, I don't think this is permanent, but I am worried about such a trajectory: your ability to win in this game has less to do with what you do during gameplay than it did before.

-------------------------------------------

In the defense of Mathwing:

I don't fully grasp it. I know it isn't perfect. I can see the limits of it's use. It has been very consistent in predicting and explaining many aspects of the game. Don't pretend it's the be all end all or that it's flawed so useless and use it for what it's meant for (somewhere in the middle) and it's great.

Similarly, most 4-5 ship imperial lists have an 85/15 match up vs. Fat Falcon. Imperials are really feeling hamstrung by what is starting to amount to 2 ships left to build with (advanced and bombers are really just not happening)... 6 named tie swarm is the exception and it's 85/15 is probably a well flown phantom (probably 80/20).

Rakky, I respect your thoughts on these boards, but in no way is this an accurate representation of what I've seen. I have very solid win rate against "Fat Falcon" lists and I always run fewer than 5 ships and don't use phantoms. I've used combinations of Interceptors, Defenders, Ties, and an occassional shuttle mixed in.

Look I think we can all agree I was right about the defender even before it came out, I said it'd be the real winner in wave four.

I may be an idiot child when it comes to math due to being disnumerate (dyslexia but for maths) but I could see it's potential once we saw the dial, even when told I was wrong because it was clearly over costed I stuck to my guns and though vindication has taken a while it's made me smile to see people discover this gem.

Just remember I liked the defender before it was cool *dons hipster glasses and a fez*

I should probably be more specific about 4-5 ship lists vs. Fat Falcon. Even after posting this one didn't quite sit right with me.

Baron Fel can tangle with anyone, fat or not. I've had mixed results with defender lists, certainly not 85/15 but not in the 60/40 land I preferred from "the good old days". Ties are everywhere and the 1 big + mini swarm of 4 ties can also tangle with most lists.

And I really meant "2 ships less to build with" not LEFT to build with (that I did fix).

I redact my posting! (but will leave it there anyway)

I do stand by 6 ships probably not being swarmy enough and interceptors certainly being capable but not if dice and skill are even. I also think of a lot of my old builds and know how they function well enough to know that skill and dice equal I'm only getting 2-3/10 wins despite my extreme experience with the lists. I am working from STARTING at this 85/15. Good piloting and dice (or really poor piloting) skew this from there. But will you enter a competitive event knowing your starting point is that low with your list.

I'm looking a bit too far into the future even for my own comfort right now; my 85/15 rant is really pointing out that a non-swarm list with more than 1-2 two dice guns is going to be pretty helpless to take down the fat falcon and soon there will be combos to make FATTER falcons that can ignore 3-4 and maybe even 5 damage a round (how that will do competitively is anyone's guess, but at the more casual level that will be devastating for new players). I see Dashing YT's also creating some nightmare match ups if your ships don't have repositioning abilities.

Thanx AlexW for checking me. I'm going to shut up for a bit and do some real world work.

If the Advanced and Defender were as overpriced as people have been saying, I should never had even close to the amount of success I've had with my Vader & Rexlar list. Is it top tier, no. But, it should not have beat the squads I did (including one Phantom + 2 Firesprays).

And the times I lost, I recognize very clearly why it was my mistake in losing.

I get really annoyed when Mathwing figures are held up as the end all, be all of X-wing game theory. Especially MajorJuggler's figures which point to an extremely broken game that needs an entire recosting.

I'm looking a bit too far into the future even for my own comfort right now; my 85/15 rant is really pointing out that a non-swarm list with more than 1-2 two dice guns is going to be pretty helpless to take down the fat falcon and soon there will be combos to make FATTER falcons that can ignore 3-4 and maybe even 5 damage a round (how that will do competitively is anyone's guess, but at the more casual level that will be devastating for new players). I see Dashing YT's also creating some nightmare match ups if your ships don't have repositioning abilities.

Thanx AlexW for checking me. I'm going to shut up for a bit and do some real world work.

I sent you a PM, too, but I do see your point about two dice guns. I have almost completely moved away from those. Though I am going back to Vader....

Yeah, it's the 2 attack ships that are going to have a lot of problems. Because of Predator, the academy Tie, which used to be a great ship, is now becoming a huge liability. To make an effective swarm you've got to sink a lot of points and lose customizability by bumping up at least to Obsidians just to keep your ships alive. I've got some hopes for 6 Tie swarms that I'll be trying out, particularly with Predator and Intimidation, but I'm afraid that 6 ships is where the effectiveness starts to tank for other reasons.