If you're now arguing that Goko Pro isn't so bad or that Isotropish should be changed or discontinued, then maybe you should do some analysis to demonstrate it. I've already downloaded the logs, parsed the games, and aggregated the results for you. Both that game data and the source code for Isotropish have been publicly available for over eight months. How much easier can I make it for you?
I'm not arguing any of that. I'm arguing that your rant which appears to boil down to "Goko Pro sucks in comparison to isotropish" is without basis. You make comments about the backwardsness of people who don't follow you, but you actually don't really give evidence for your position. In fact, there isn't evidence for your position except for anecdotal "it seems better" evidence. Admittedly, I don't have any better evidence the other way, and I'm not claiming that I do. My point is to get off your high horse.
It's actually very difficult for me to do analysis on the accuracy of isotropish, because in order to measure it accurately, I need to constantly scrape your board, since what is meaningful is the prediction AT THE TIME THE GAME IS PLAYED, and since you continuously update, you don't keep records to cross-reference. So it's actually really complicated. On the other hand, it should be relatively easy to measure accuracy at the time you rate a game, since you have to pull all the pertinent information anyway.
Of course, this isn't going to help in a comparison against the Pro mode ratings, which don't give predictions at all.
Goko Pro is a disaster because it's TrueSkill that's been deliberately broken.
This is just false. I don't know why you have some kind of strange fascination that rating systems having these kinds of features are somehow derivatives of TrueSkill. TS isn't the first rating system to incorporate these features, it isn't the best-known, it doesn't have the best reputation. You can say that their system is a derivative of Glicko (as is TS), or of Elo, or you can be most accurate and say it's their own system.
In any case, the point you're making here isn't relevant - Basically what you are saying is "It's TS except it's been changed", which would be bad if TS were perfect, but it isn't. So we don't actually know whether the changes are good or bad, which means that just because it's been changed doesn't mean it's worse.
I haven't gone out of my way to prove that Goko Pro sucks
Rating system has changed since this comment
because everyone already knows it sucks.
Again, I'm not saying that the Pro rankings are good. But most of these problems are entirely shared by isotropish "level". I can make a similarly scathing attack on a bunch of problems with TS (many of them will be the same attacks, actually). But yeah, there are definite problems with the current Pro ratings, nobody is arguing that.
Nobody likes it. Nobody defends it.
This is both false and irrelevant
One forum member even went so far as to implement his own proprietary system as an alternative.
I didn't know that! Thanks for informing me!
Isotropish isn't meant to be the One True Dominion Rating System. It's just standard TrueSkill with standard parameter values.
Oh, and one last thing here. Two of the parameters are set to their standard values (the 25 and (25/3)^2. But there are other parameters at play - how were they set? Doesn't "standard" TS have a time decay that isotropish doesn't? In any case, whether they're standard or not doesn't mean they're good.
But again, this isn't the point. The point isn't that your system is worse than the Pro system. The point is to not trumpet your superiority when you have no evidence of your superiority.