18 hours ago, MajorJuggler said:

I am not a playtester or paid by FFG, but I have the demonstrated capability (at present seemingly uniquely) to predict relevant point costs far better than FFG's development and playtesting process. I can do things that nobody in their entire company seems to be anywhere close to doing

Dude... I respect the stuff you've contributed over the years but you're going *way* over the line here, and I don't just mean in politeness. If you're going to take off the gloves with personal attacks I'm going to have to bring you back down to earth: a statement like that makes it clear that your ego has vastly exceeded your accomplishments at this point.

Rather than get into a psychic-like pseudoscience mud slinging about how many things you have "correctly predicted" vs. mispredicted vs. "justified after the fact - with math!", let me just address the fundamentals here: things like "jousting efficient" values had some amount of utility in the wave 1 days where we lined up with 2-4 ship types and k-turned around each other and rolled dice, but we've come a long way since then and the amount of assumptions that you need to make to derive a single "goodness value" out of a modern game of x-wing makes the results completely non-general. I'll use your parattanni analysis as an example: coming up with a set of assumptions that makes that squad come out with high numbers *after we already know it is good* is neither difficult, nor impressive. To prove a model you need to make consistently accurate *predictions*, or find disproportionately good game interactions that are not obvious to the general population of players. i.e. if your model was highly accurate "in general" (if such a thing even exists - more on that later) you should have been able to have it *discover* new and broken squads and interactions and then *prove* those by winning tournaments with them before anyone else even considered it.

Don't get me wrong, I'm not being "anti-math" here. Obviously I write a tool that allows people to work out good combinations and probabilities and such in specific game situations that I feel is fairly valuable both in squad building and in game decision making. But don't miss the key point there: *in specific game situations*. While it would be technically easy to have the tool simulate tons of combinations of dice rolling for different squads or cards or builds, the number of assumptions required to try and claim "therefore math says this one is 10% better" would be intellectually dishonest.

The reality is that at this point the game is complex enough that you can't just derive an "awesomeness number" for everything. Indeed to that end, List Juggler and metawing are the more relevant tools to see what is working well and what isn't (and I thank you for your continued technical contributions to those!). Thus if the goal is to derive the "value" of certain squads and upgrades and so on, I think analyzing that data set is ultimately the way to go. Of course you need to make guesses before releasing new content and some amount of math can inform those guesses, but I absolutely reject your claim that you've "demonstrated unique capability to predict relevant point costs far better than FFG".

Anyways I don't want to get personal here, but you opened the door yourself by making it clear that you need to take a serious look at your own ego here. And I'll remind you in advance that you aren't the only one with math degrees and jobs and technical experience here, in case you try to play the appeal to authority card again

If you want to make your own point costs mod or whatever, please go right ahead. But if you want to maintain credibility here maybe do it without the associated ego trip and mud slinging at FFG.

17 hours ago, sinclair5150 said:

The sheer amount of hubris in this post is staggering.

XWing2 literally incorporates a huge number of ideas that were first presented by MJ. At _worst_ this means that their entire dev team is coming up with the same ideas that MJ is coming up with on his own. At best it means they're already letting him do a bunch of grunt work, waiting for him to publish it, then incorporating it.

Either way, it's not Hubris when he's right.

Option 4. Make a generic space flight system game. Put it on Kickstarter to cover your material costs, make sure you raise enough funds to have app access.

I got quite a bit of mileage out of your house rules the first year. I'd certainly be interested in a flight system game designed by you. I'm not sure I'll ever get into another "living/customizable" game, but if you balanced something out and where able to deliver a complete set with 5 or 6 completed ships per 2 or 3 factions... yeah, I'd bite.

I would be happy if u just balanced 1.0 . Now that it's over u don't have to future proof it I just hope u keep it as simple as possible - maybe just bans and point cost adjustment

We can assume the last two or three waves were some kind of sick joke. So no poons no bomb buffs no sheathapede

I kid, I kid.

1 hour ago, Punning Pundit said:

XWing2 literally incorporates a huge number of ideas that were first presented by MJ. At _worst_ this means that their entire dev team is coming up with the same ideas that MJ is coming up with on his own. At best it means they're already letting him do a bunch of grunt work, waiting for him to publish it, then incorporating it.

Either way, it's not Hubris when he's right.

Evidence they took Juggler's ideas?

18 minutes ago, mightyspacepope said:

Evidence they took Juggler's ideas?

Don't even go down this road man... it's like psychic's proving their "predictions" are right. It's trivial to cherry pick, warp or reinterpret any number of things with sufficiently many posts and no amount of arguing back and forth will convince anyone to change their minds. Note that I'm specifically not making a statement on whether FFG did or did not take ideas from the community (it's fair to assume that's one of many sources of their inputs), but that the effort and value of constructing something of the scope of 2.0 is far higher than throwing out suggestions on the forums, even if backed up by math.

I'll point out the converse though: some of the most interesting things in 2.0 - linked actions, red actions and such - I have not seen suggested in that form, and yet retrospectively they seem very clever and almost "obvious" design. We all know PTL and associated action/token stacking is dumb, but I've not seen anyone suggest a solution that is nearly as elegant as what they came up with for 2.0, so good on them!

And that leads to the point: arguing about "who gets credit" on forums is not going to accomplish anything useful, and it's even very well defined in the first place. If people want to create their own unique rule sets, costs, or even entire games that should absolutely be encouraged, but nothing in that process should involve personal attacks against FFG, ego trips or other degenerate behavior.

21 hours ago, MajorJuggler said:

So I have a few options:

1. Inquire about consulting for them, since they still have a clear need to polish balance before launch. (This will almost certainly result in them politely saying they don't need me, and then they will turn around and just use my ideas anyway.)

3 hours ago, mightyspacepope said:

Evidence?

The FFG designers have already asked me for these numbers before.

3 hours ago, punkUser said:

Rather than get into a psychic-like pseudoscience mud slinging about how many things you have "correctly predicted" vs. mispredicted vs. "justified after the fact - with math!", let me just address the fundamentals here: things like "jousting efficient" values had some amount of utility in the wave 1 days where we lined up with 2-4 ship types and k-turned around each other and rolled dice, but we've come a long way since then and the amount of assumptions that you need to make to derive a single "goodness value" out of a modern game of x-wing makes the results completely non-general.

The general answer to this exists earlier upthread:

5 hours ago, MajorJuggler said:

There is a common misconception that jousting values are not useful for turrets or arc dodgers because they don't spend most of their time in straight line jousts.

Jousting efficiency, firing duty cycle, and total effective efficiency are all related; if you know two of the variables then you can determine the remaining one. This lets you take a turret's straight-line efficiency, and determine how much better its fire duty cycle needs to be relative to a straight line jouster to "break even". Above this duty cycle the turret is better; below this duty cycle the jouster is better. You can also directly modify the jousting value by folding the firing duty cycle directly into the ship's expected damage output, thereby directly calculating the total effective efficiency.

The process for dual action reposition aces is essentially the same, except there are multiple figures of merit depending on how many actions an ace is burning for repositioning on any given turn.

Through analytical playtesting and historical data, the cost efficiencies that turrets and arc dodgers should have to put them in the "goldilocks zone" is fairly well established.

3 hours ago, punkUser said:

I'll use your parattanni analysis as an example: coming up with a set of assumptions that makes that squad come out with high numbers *after we already know it is good* is neither difficult, nor impressive. To prove a model you need to make consistently accurate *predictions*, or find disproportionately good game interactions that are not obvious to the general population of players. i.e. if your model was highly accurate "in general" (if such a thing even exists - more on that later) you should have been able to have it *discover* new and broken squads and interactions and then *prove* those by winning tournaments with them before anyone else even considered it.

You make an excellent point to making a distinction between discovering combinations, and analyzing and making predictions of those combinations. There is another separate discussion about the philosophy of playtesting; my philosophy is for the playtesters to discover, and the developers to analyze. You're right that a mathematical model on it's own can only process what you feed into it. The flip side, is that even just evaluating the "basic configurations" of new FFG content has historically revealed significant pricing inconsistencies.

As for challenging the model's ability to make predictions, here are a few examples:

• The B-wing would overshadow the X-wing as a jouster. Technically this was after release, but I made this prediction well before it was commonly accepted as fact, or was clearly showing up in the tournament results.
• The Z-95 would replace both the X-wing and B-wing as the preferred jouster
• I pronounced the stock Defender dead on arrival. I correctly called this before the dial was revealed. Several playtesters vigorously and publicly debated that I couldn't pronounce it DOA because I hadn't seen the dial. They were wrong, because the math was clear that no dial existed that could possibly make it cost efficient. It had a lower cost efficiency than turreted ships.
• TLT Y-wings would displace Z-95's and non-crackshot TIE Fighters as the preferred cost efficient filler. [edit: I had a private conversation with the playtester who was calculating the TLT numbers for FFG when they were designing TLT. He was convinced that TLT would be fine. I disagreed.]
• I "invented" Commonwealth Defenders as soon as the Imperial Veterans preview came out. Upon release it immediately won multiple nationals.
• I quantifiably predicted why Vader + ATC was going to be very strong, but the rest of the TIE Advanced would be "meh". [edit: I did make an entire post dedicated to generics with accuracy corrector being a dangerous design precedent for any ATT2 ship with a system slot, but then hyper-defense quickly outpaced the ability for 2 guaranteed hits to do meaningful damage.]
• The Inquisitor was predicted as obviously being too strong. (>110% jousting efficiency on a dual action PS8 ace leads to an easy conclusion)
• I didn't invent Parattanni, but I privately ran the numbers before it really "caught on". I was discussing a similar list on a S&V thread when Old Para, the originator of Parattanni, pointed it out in public for the first time. I was about to go to Regionals, so ran some numbers and kept it private for competitive advantage. Then I went 7-1 at Regionals (Top 4, still kicking myself I didn't win, my own stupid mistake) and used it and some post manaroo nerf variants to get into Deep Core for a couple seasons in the vassal league.
• I predicted that PS11 Rebel Fenn + Hotshot co-pilot was an absurdly strong piece. (Again, >100% efficiency on PS11 leads to easy conclusions).

There's probably more that I can't remember offhand, these are kind of just the bigger ones. There is a lot of radio silence in there, because I have usually not done the analysis pre-release. In cases of post-release evaluation (which are too numerous to list), the models can be considered "explanatory" rather than "predictive", but the analysis process is the same.

3 hours ago, punkUser said:

Don't get me wrong, I'm not being "anti-math" here. Obviously I write a tool that allows people to work out good combinations and probabilities and such in specific game situations that I feel is fairly valuable both in squad building and in game decision making. But don't miss the key point there: *in specific game situations*. While it would be technically easy to have the tool simulate tons of combinations of dice rolling for different squads or cards or builds, the number of assumptions required to try and claim "therefore math says this one is 10% better" would be intellectually dishonest.

The reality is that at this point the game is complex enough that you can't just derive an "awesomeness number" for everything. Indeed to that end, List Juggler and metawing are the more relevant tools to see what is working well and what isn't (and I thank you for your continued technical contributions to those!). Thus if the goal is to derive the "value" of certain squads and upgrades and so on, I think analyzing that data set is ultimately the way to go. Of course you need to make guesses before releasing new content and some amount of math can inform those guesses, but I absolutely reject your claim that you've "demonstrated unique capability to predict relevant point costs far better than FFG".

This is a very good starting point for a very interesting discussion which is not yet 100% solved, and has a very long and technical answer. :-)

The important takeaway is to recognize that any model has limitations, and to properly recognize it's limitations. In this case the TL;DR is that the model is sufficient to point out some of FFG's costing mistakes with a very high degree of confidence.

3 hours ago, punkUser said:

Anyways I don't want to get personal here, but you opened the door yourself by making it clear that you need to take a serious look at your own ego here. And I'll remind you in advance that you aren't the only one with math degrees and jobs and technical experience here, in case you try to play the appeal to authority card again

No worries, and just as you don't intend for it to be personal neither do I. In the end we are all pushing around plastic spaceships! When I "take the gloves off" I am speaking broadly about FFG's capabilities as a company, and the industry as a whole. I will be the first to sing the praises of many of their designers, they generally do an excellent job at design and development. But technical balance is an entirely different ball of wax. They are just not geared up to do it right, because it's not a business priority for them.

I am kind of surprised that after 4-ish years I am still the only MathWing guy that has built and disclosed any sort of public costing model. I would have expected to be having conversations with at least a few other people by now about hyperspace convergence (not the star wars kind) of meta-dependent action economies. :-P

4 minutes ago, MajorJuggler said:

You make an excellent point to making a distinction between discovering combinations, and analyzing and making predictions of those combinations. There is another separate discussion about the philosophy of playtesting; my philosophy is for the playtesters to discover, and the developers to analyze. You're right that a mathematical model on it's own can only process what you feed into it. The flip side, is that even just evaluating the "basic configurations" of new FFG content has historically revealed significant pricing inconsistencies.

Sure, you can absolutely find outliers and direct play-testing based on numerical analysis, but I would still argue that playtesting must trump the analysis in any non-trivial cases, simply due to the number of assumptions and thus significant error bars that any analysis of a game of the complexity of x-wing entails.

4 minutes ago, MajorJuggler said:

As for challenging the model's ability to make predictions, here are a few examples:

As I mentioned above, while I have respect for you contributing your various analyses to the community, don't let confirmation bias lull you into a false sense of superiority over the designers or anyone else. It's the attitude of your original post that I take issue with, not the fact that you're trying to apply analysis to the game (something that a fair number of us do in a variety of ways and I am a proponent for!).

4 minutes ago, MajorJuggler said:

The important takeaway is to recognize that any model has limitations, and to properly recognize it's limitations.

Absolutely, and this is fundamentally the point that I think people need to keep in mind. A model is only as good as the assumptions you are making. I personally question the utility of some of the models you have posted about in the past due to how many assumptions are required (we can get more into this if you'd like), but I certainly have no issue with you or anyone going down the rabbit hole as long as you make both the details and the limitations clear to people reading it. Ex. it's very hard to imagine a model that would cost something like Blair's 7 z's against nym/miranda appropriately in all matchups, and thus I'd be skeptical of anyone claiming to do that without way more information about all of the assumptions.

And really, if you carry the error bars through your entire analysis for each of the assumptions you make, they grow so large as to be meaningless very quickly.

... and that's a good thing, otherwise we'd just "solve" x-wing analytically and there'd be no point in playing

Anyways I appreciate the more thoughtful reply. Let's just keep things in perspective and civil, particularly when talking about the designers, who are "real people too"

1 minute ago, punkUser said:

As I mentioned above, while I have respect for you contributing your various analyses to the community, don't let confirmation bias lull you into a false sense of superiority over the designers or anyone else. It's the attitude of your original post that I take issue with, not the fact that you're trying to apply analysis to the game (something that a fair number of us do in a variety of ways and I am a proponent for!).

You didn't seem to think that the model could actually be used as a predictor, so I provided a list! It's a pretty long one with a lot of heavy hitters in it, so if that's not a convincing enough list, I'm not really sure what would be. Which is fine, I don't care if people believe it or not, but you asked if it existed, so, there it is. :-)

To show confirmation bias you would have to demonstrate examples of where I made predictions that turned out to be wrong, to show that I am cherry picking. Every time I bring this up the conversation suddenly ends or changes topics... which is a little unfortunate because I'm sure I'm not always 100% right, and feedback is super helpful!

4 minutes ago, MajorJuggler said:

You didn't seem to think that the model could actually be used as a predictor, so I provided a list! It's a pretty long one with a lot of heavy hitters in it, so if that's not a convincing enough list, I'm not really sure what would be. Which is fine, I don't care if people believe it or not, but you asked if it existed, so, there it is. ?

To show confirmation bias you would have to demonstrate examples of where I made predictions that turned out to be wrong, to show that I am cherry picking. Every time I bring this up the conversation suddenly ends or changes topics... which is a little unfortunate because I'm sure I'm not always 100% right, and feedback is super helpful!

Not for nuthin' but you also called out the stresshog (albeit with the ion cannon) before Heaver shocked everybody with it on the way to a world title.

21 minutes ago, punkUser said:

And really, if you carry the error bars through your entire analysis for each of the assumptions you make, they grow so large as to be meaningless very quickly.

This is a very strong quantitative claim that you have made only qualitatively!

Actually, because reality (roughly) follows the inverse square law, the opposite ends up being true: it's virtually impossible for mathematical costing estimates to be way out in left field. Unless your data analytics and figures of merit are REALLY out of whack, the spread of ship competitiveness will be much tighter than what FFG's designs have come out of the gate as. They haven't caught up to MathWing 1.0 yet!

1 minute ago, MajorJuggler said:

You didn't seem to think that the model could actually be used as a predictor, so I provided a list! It's a pretty long one with a lot of heavy hitters in it, so if that's not a convincing enough list, I'm not really sure what would be. Which is fine, I don't care if people believe it or not, but you asked if it existed, so, there it is. ?

My point above stands; if you're familiar with critical thinking you know this line discussion is exactly the same one that psychics use and not one I really want to waste a lot of time on for the reasons I mentioned Let's instead stick to the technical argument that your list of assumptions is too long to provide meaningful generalizations in most cases. If you want to get more specific on something in "MathWing 3.0" you'll need to give more detail. But you've mentioned publishing a paper, so that'd be a better and more direct way to get some peer review

1 minute ago, MajorJuggler said:

To show confirmation bias you would have to demonstrate examples of where I made predictions that turned out to be wrong, to show that I am cherry picking.

Without spending a ton of time on this, let's just pick the cost reduction of sensor jammer in your community balance tweaks as a pertinent example for my point here. While it makes some amount of sense under the assumptions you were making to compute its value, look at Ghost/Fenn... clearly it's worth at least 4 points if not more value there! And that's ultimately my point: there are too many interactions these days to just rely on averages of chasses, fire duty cycles, attack/defense results and mods and claim they have meaningful predictive power outside of those specific assumptions and average cases. If you were to model the full intervals/error bars through all of these assumptions I think you are probably aware that they immediately become nutty.

These sorts of analyses can obviously have utility in specific cases. Again I'll point to the fact that I've done some work on that front myself and will continue to And certainly it can and should *inform* design. But when you stray too far from the original assumptions in terms of trying to claim general results (i.e. anything that assigns a single value to a squad frankly), you're overstepping the math. And it gets worse if you then try and confuse non-experts with technical terms.

Did you look at the other changes to Ghost / Fenn?

:-D

9 minutes ago, MajorJuggler said:

Actually, because reality (roughly) follows the inverse square law, the opposite ends up being true: it's virtually impossible for mathematical costing estimates to be way out in left field. Unless your data analytics and figures of merit are REALLY out of whack, the spread of ship competitiveness will be much tighter than what FFG's designs have come out of the gate as. They haven't caught up to MathWing 1.0 yet!

Come on dude... I'm not impressed by your meaningless vomiting of technical terminology. "Reality follows the inverse square law"... "virtually impossible for mathematical costing estimates to be way out in left field"... really? Trivially there are way more cost functions that would be "way out in left field" than would approach any theoretically "correct" model. And fundamentally you'd have to even define what you mean by a "correct" model here as there are a variety of interesting functions one might want to - often simultaneously - optimize. "Ship competitiveness" for instance is trivially satisfied by just having a single ship after all.

3 minutes ago, MajorJuggler said:

Did you look at the other changes to Ghost / Fenn?

I'm not sure if you actually missed my point or if you're agreeing with it? It's a perfect example of where costing something based on averages and layers of assumptions is not particularly useful, which is precisely what I've been saying this whole time. And it's an example of a case where it seems reasonable based on a mathematical model but everyone will trivially agree that it falls apart with other interactions (i.e. as it has in the current meta).

One can certainly argue that avoiding too many interactions like that is probably a good plan if you are designing a game from scratch, and indeed they seem to have taken this to heart in a number of cases in 2.0, which is great! But given that there are a myriad of those sorts of interactions in 1.0 today, I still question the generality of a model that by necessity just averages a lot of it out.

The intent of the OP was not to get into a debate about the veracity of the model.

Inevitably there are always some individuals that vehemently deny what is possible, and usually it ends up coming across as personal. Then later I make some new prediction that comes true, and then different people come out of the woodwork to say I haven't got a clue, and the cycle continues.

This has been a constant theme for the last 4 years, and seems to happen anytime I say anything about math at this point, so, I'm kind of used to it and desensitized to it by this point.

15 minutes ago, punkUser said:

Come on dude... I'm not impressed by your meaningless vomiting of technical terminology.

So, I will just respond in kind that everything you have stated in this thread is also a meaningless vomiting of technical terminology. That sword cuts both ways. ;-)

There. Neither of us have said anything, so lets move on and stay on topic. :-)

1 minute ago, MajorJuggler said:

The intent of the OP was not to get into a debate about the veracity of the model.

Right, and if you had kept the ego out of it, I would have happily stayed away But you made a bold claim about how much better you were than everyone else and I don't think you should be surprised at this point if that garners some disagreement, particularly since you say that you're "used to it" at this point. I'll humbly suggest that in the future if you don't want to derail your threads, stick to the basic questions and keep the grandstanding out of it.

1 minute ago, MajorJuggler said:

Inevitably there are always some individuals that vehemently deny what is possible, and usually it ends up coming across as personal. Then later I make some new prediction that comes true, and then different people come out of the woodwork to say I haven't got a clue, and the cycle continues.

Just to avoid the potential strawman for others reading this thread, I would never claim that you "haven't got a clue". Indeed my claim is just that you sometimes over-generalize the applicability of your results and occasionally adapt the model to fit the meta rather than the other way around.

If you're fully aware of the limitations of your methods that's great; I would just ask that you be more up front with people about that rather than basically claiming no one at FFG or elsewhere have the technical skills to do what you have done. In reality it's more a difference in opinion on the utility of it beyond a certain threshold of assumptions than the technical capability.

1 minute ago, MajorJuggler said:

So, I will just respond in kind that everything you have stated in this thread is also a meaningless vomiting of technical terminology. That sword cuts both ways. ?

I've attempted to avoid trying to impress you with my technical qualifications here because in the end what matters is the actual content of the arguments. I'm merely suggesting that for the specific case I quoted if your statements don't make a lot of sense to someone with a couple math degrees then maybe you should reconsider how you explain things, particular to a general audience on the forums for a tabletop game

Anyways I agree we've both said everything that is useful at this point. If you want to have a more peer-review style of your methods at some point, I'm totally interested in seeing the upcoming paper. As far as this thread goes, I'll just reiterate for the record that my primary point is just that you should tone down the ego a bit in the future

1 hour ago, punkUser said:

... and that's a good thing, otherwise we'd just "solve" x-wing analytically and there'd be no point in playing

Back on the original topic, this is actually why I went down the path of making Community Mod. X-wing has already been solved analytically for myself and many other players. And we are bored so we have stopped playing.

There's a few lists that are at the top of the competitive food chain, and it is very clear why they are there. Having only a few lists to play against gets boring pretty fast, especially when those lists are inherently "non-fun" game experiences. I'm not sure if you are aware, but I have stopped playing locally in person for precisely the reason that you quoted above. It's been solved, and it's unbalanced, and... just not fun.

Now, if there were a bunch of different types of squads that I could fly and not immediately get kicked to the curb, then that's another story. From a player's perspective I wouldn't care if the game is "solved", because there's sufficient variety to keep things fresh. Instead of 3-5 viable lists there could be 100-200. That will keep me occupied for a while. :-)

My plan for X-wing 2.0, from a competitive standpoint, is to privately analyze wave 1 and take advantage of imbalances and bring the overpowered stuff before everyone else figures it out and FFG nerfes it. Well, I'll do that for a little while anyway, then I'll want to fly a bigger variety of content, and hopefully FFG has caught up with their own costing by then so a wider variety of squads are viable.

I would rather the game be better balanced out of the gate, that would be more fun to me, even if it was not in an official format. The OP is asking if other people would be interested in the same.

Wait ! You are telling us that as soon as the players lists are revealed, the winner is known and there is no point in playing ?

Ho man ! You are right ! X-wing (1 and 2) needs to be fixed !1!

I will vouch for MajorJuggler. He was right about many things and has contributed fairly to the community in many ways. Grand statements, but I agree with them.

I'd be interested in a balance mod, because - I fear that at the rate they're planning to push the product - proper playtesting and balance issues have not been a consideration.

6 minutes ago, Keffisch said:

I'd be interested in a balance mod, because - I fear that at the rate they're planning to push the product - proper playtesting and balance issues have not been a consideration.

It's not because they announced it 2 days ago that they were not working on it for a very long time.

16 minutes ago, Cartchan said:

It's not because they announced it 2 days ago that they were not working on it for a very long time.

No?

Would you care to elaborate?

3 minutes ago, Keffisch said:

No?

Would you care to elaborate?

2.0 has been in development for over a year.

1 hour ago, citruscannon said:

2.0 has been in development for over a year.

And hopefully it shows in the design of the core rules.

But that has very little to do with the speed with which they are going to release future waves - not to mention the quality of playtesting and balance (i.e. powercreep). going forward.