jrh_cardinal wrote:You guys wonder why I got tired of arguing. Every post of mine until the last two only dealt with the issues and trying to prove your viewpoints wrong. All along the way max avoids my points or attempts to make excuses why my points aren't valid rather than actually trying to prove them wrong, then starts going after me personally.
I never went after you personally, because unlike you, I'm not getting worked up about this debate. I want this change to be implemented, but only because I think it rationally makes sense, not because I care a lot about the ratings system.
Now you do the same thing. Honestly, you guys are unbelievable. Can't you just have a heated discussion about the issues, without bringing in the integrity of the two people involved?
Not once did I question your integrity, and it's unfortunate that you would even suggest that I have, as a way to indicate that somehow I didn't debate you honestly. The only remark I made that was even somewhat unrelated to the content of the debate itself was when I said it was childish to paraphrase my comment into something I didn't say, and then attack me for it. And that
was an unfair method of debate, and I called you on it. That's not the same thing as an
ad hominem attack.
I understand that you're
saying that your points show why this system would be worse than the current one, but in reality that's incorrect. There's only one argument you've consistently made that you've even suggested would make the system worse, which is that if everyone is within 0.1 of the mean, then it's harder to figure out who to play with. The reason why this is totally irrelevant is because that's true now. Most people fall around 4.7 plus or minus 0.1, with a smaller minority at 4.6 and 4.9. Thus even if most players did fall in the range you quoted, it would be the same situation as now - if you're trying to discriminate and find players that are well above average, then all you have to do is find the ones which are above 0.1.
At any rate, in a normal distribution, roughly 2/3 of the players would be within one SD of the mean; you've indicated that you understand statistics, so this shouldn't come as a surprise to you. The problem is that we have no idea what the standard deviation will be until we implement this system. We
could probably estimate it if we had all the information on the current ratings system, which we do not. If I had to guess, I would imagine the standard deviation would be 0.2 under this system, not 0.1 (because I think that people will rate at a slightly higher rate). If that's the case, it would be even easier to discriminate than it is currently. So either it gets better or it stays the same. That's why there's no logical reason not to do this.