Gyrok, on 01 June 2016 - 09:54 AM, said:
Actually, the genetic fallacy is true here.
The argument is that the skill level those players possess makes their opinion more valuable than other opinions not coming from them.
By default, that argument is set to assume that the players who are not as skilled at the game have less fundamental understanding.
No that is not the assumption. I explained why.
It's true that anyone with a good head for game balance can look at the best players results and analyze them, what you can't do is use bad players results. Good players are more likely to be the ones doing this, because they are producing these results themselves and are directly invested in the competitive meta. Professional coaches, good referees, experienced tournament organizers and others with incentive to invest in understanding the meta are also more likely to make good predictions.
Beyond that you have to look at the actual analysis to determine if it's good, my assertion is that Twinky etc has made a good judgement about the 3, I think they are correct about it and that this will show in the upcoming tournament results.
I assume you think they are wrong and that the tournament will prove them wrong, so we'll actually be able to wait for the experiment to be done and revisit this after.
Quote
Yes, knowing the winning moves is all you need to know.

Choosing mech and build is one of those moves, thus you need to be very good at evaluating the strenght of mechs to be competitive.
Thanks for proving my point.
Quote
Yes, but it can not refute the strength of a move by using examples of bad results. That is one the things where anecdotal evidence is useless, and quite frankly any evidence short of shooting down the credibility of the demonstration itself which you try to do here:
Quote
This assertion that the demonstrated peaks are against so called "potatoes" etc definitely requires more than anecdotal evidence to be believable. Again an example where anecdotes would be useless since the claim is that of a general trend.
It's just as likely that the increased player count temporarily improved matchmaking, as matchmaking quality requires many players, and therefore it's entirely possible that Twinky et al even met stronger resistance than ususal.
I'm not claiming either of these scenarios is correct, just demonstrating that you can't speculate your way to knowing it. I think in absence of knowing the basic position is that the opposition was pretty much as usual.
We do know that the matches was full of assaults, but as has been pointed out the kill efficiency has been just as good as ususal among top kdk 3 users, there is no really good basis for claiming that the results are "bloated"
Quote
Did I not just explain why anecdotal evidence is useful to demonstrate peak performance but not useful to demonstrate low performance?
Yes I did, you are now trying to use anecdotal evidence for one of those things it is actually useless for.
As with the example of the "valle CC" streetfighter move, anecdotal evidence is plenty enough to demonstrate a new peak. Conversely a player trying to use the Valle CC and failing would not be evidence against it's strength, it would just be a bad player.
I understand that it feels unfair that something is useful only for the other side of the discussion, but it's true.
Quote
No it doesn't, as I just explained subpar performance can not be useful in anecdotal evidence. Only different kinds of peaks or cases that refute a general absolute assumption are useful.
Quote
Since the assumption that the Kodiak is the strongest assault does not require that everyone must perform well with it, even the most OP mech imaginable would have people fail in it, it cannot be refuted by example of bad results or failure to realize it's potential. If the assertion was for example "The KDK-3 cannot lose" it would be refuted by examples of bad performance, that is probably why people like to strawman balance concerns that way. You can see examples of that on the forums, the typical strawman phrase is "it's not a win button", seeing that kind of phrase is a certain way to spot a useless post.
At best you can use that kind of data to demonstrate that a high performance peak move also has a high or low skill threshold, but the actual strength of the move must still be judged by it's highest consistently achievable peak, in other words what can the best players consistently do with it.
The truth is that no amount of demonstrated bad results in the Kodiak would refute the demonstrated peak, not even a thousand or a billion screens could do it.
The peak has been demonstrated in both height and consistency.
Quote
That argument completely refutes any attempt to justify the argument that personal opinion about the mech being too strong is valid.
That doesn't actually matter and it doesn't refute anything. The best demonstrated results and the best possible analysis of those results are the relevant thing.
If someone good is much better at ballistics than lasers the only consequence of that is that his results in ballistic mechs will be useful in balance discussion and his results in laser mechs will not.
Obviously it's a very good idea to also balance things across different skill brackets, because you don't want imbalances among weak players to make the game to hard to get into, but you have to start at the top and balance your way down without compromising the top balance. The top is easier to analyze since it is only at the of the skill curve that it is easy/possible to isolate away skill as a factor.
Edited by Sjorpha, 01 June 2016 - 02:31 PM.