Average opponent rating...

Originally posted by ItalyBoyBlue
Doctor scribbles,

Most of what you have posted is both incoherent and immature. There is no drawback to having this statistic posted despite the fact that it is obvious you are unable to comprehend its usefulness in showing rating inflation/surpression (which can and does exist). I think it'd be best if you abstained in posting your wayward, irrelevent insights. They really aren't helping anybody.

As requested by me before, could you provide the maths as to how this statistic would be of any use whatsoever? Show us what percentage of games a 1500 would have to win against exclusively 1200s to maintain 1500 rating, and what percentage of games a 1500 needs to win against exclusively 1800s to maintain a 1500 rating.

Also, can you explain how this statistic would aid in showing rating suppression? If I wanted to suppress my rating having previously played all 1800s, all I'd have to do is throw 4 games against 1400s and I'd be down 136 points. Now, considering I may have played 300 games against 1800s, my average opponent rating would still show at about 1800.

Can you show how it would aid in showing rating inflation? If I only exlusively play players well below my ability, then I have to win all of the games, or my rating is going to take a huge hit. Work it out.

You do also realise that working this statistic out is a fairly hefty whack on the servers, for what you have yet to show would have any use.

Seeing as we seem to have to validate our maths knowledge to contribute to this thread, I have a Computer Engineering Degree which was heavily mathematical.

On another point, I recommend you get ouroboros' script to hide players ratings on the game screen, as it seems you want to not only play an opponent's rating, but your opponents opponents' average ratings. IMHO, its a very bad idea to play the rating as opposed to the board. The best move is the best move is the best move.

And seeing as you are attacking DS' immaturity (where you may have a point), he is not the user who has thrown a serious hissy-fit in this thread, resulting in numerous posts being modded for profanities, nor is he the one who has started shouting when people say their idea is crap.

If you really are a maths whizz, then all I can do is recommend you calm down a bit, stop shouting, stop swearing and start explaining mathematically why your idea is so great, as presently, you are way too emotional and are making absolutely no sense, not least because your reason for needing this stat keeps changing.

Do the maths, get back to me.

D

Here's some info on ELO.

http://en.wikipedia.org/wiki/Elo_rating#Selective_pairing

Originally posted by AThousandYoung
Here's some info on ELO.

http://en.wikipedia.org/wiki/Elo_rating#Selective_pairing

First, what are your credentials?

Second, the idea proposed in this thread does nothing to alleviate the issues raised in that article, because if you're going to to assume that your opponent might artificially manipulate his rating, then you must also assume that your opponent's past opponents could have artificially manipulated theirs as well, so averaging those ratings doesn't yield any extra accuracy in assessing your opponent's strength. That is, given two players, one of whom has artificially manipulated his rating, you would not be able to conclude which is which based on an average of their past opponents' ratings, any of which may also have been manipulated.

In short, if you don't trust your opponent's rating, why put any more trust in the ratings of his opponents? That is all the proposed idea boils down to, with respect to the issues surrounding the sort of manipulation indicated in the article.

And I don't think that was even the original motivation behind this Site Idea. The original motivation was rooted in the bogus idea that a 1400 player who has played only 1200 opponents is weaker than a 1400 player who has played only 1800 opponents, simply in virtue of having played weaker opponents. Even if this were intentional, this is not the sort of selective pairing referred to in the article. The Elo system accounts for this sort just fine (that's the essential feature of the system!) - those two 1400 players do in fact expect even equity, barring any artificial manipulation of the sort described in the article.

Originally posted by DoctorScribbles
ItalyBoyBlue,

You are a retard. I challenge you to an objectively judged math contest.

Doctor Scribbles,

You are very immature. I don't see how a "math contest" between us would solve anything. Also, I am rubber and you are glue... or however that goes - you obviously still have a 5th grader's mentality so I'm sure you remember that comeback vividly. Also since your argument above directly contradicts the logic I have laid out, why not try explaining yourself the next time you inevitably repeat yourself. Do you believe it is better practice to be mildly successful playing 1800 rated players or to be wildly successful playing 1200 players??? You honestly think it's the same thing??? Capablanca disagrees with you, if I must I will quote advice from one of the many of his books I own.

Everyone else,

The point I was trying to make is that the mathematics used in the ELO ratings are insufficient in describing players' abilities in certain cases. If you don't understand and ELO ratings are enough for you, well thats just fine. This string really doesn't need your input. This is for the rest of us that do understand what is going on. I don't see how having that statistic provided would be detrimental in any way, especially since the information is already provided in the game history anyways. It would just be nice to have it indexed as a matter of conveinence.

Originally posted by ItalyBoyBlue
This is for the rest of us that do understand what is going on.

How many do you count as being in the know? By my count there are only two people who support your idea, one of whom explicitly said he was doing it just to humor you out of pity because your idea is harmless. And the other guy sounds like some idiot from Missouri, a region whose low intellect is surpassed only by Italy and a few small islands in the South Pacific.

Originally posted by DoctorScribbles
How many do you count as in the know? By my count there are only two people who support your idea, one of whom explicitly said he was doing it just to humor you because your idea is harmless. And the other guy sounds like some idiot from Missouri.

Not everyone who reads these strings posts. It really doesn't matter anyways. I am right, and you consistently criticize me without explaining yourself. WHY DO YOU CARE SO MUCH???

Ooooooh and nice comment about Italy, racist. When's the last time Scribblesland conquered the known world and more or less have dominated the artworld for most of the last 2000+ years?

Originally posted by ItalyBoyBlue
WHY DO YOU CARE SO MUCH???

As I said before, granting your request would only serve to reinforce an ignorant misconception about an essential feature of the game.

Originally posted by DoctorScribbles
As I said before, granting your request would only serve to reinforce an ignorant misconception about an essential feature of the game.

I'm afraid you're the one suffering from a severe misconception. And ELO ratings are by definition not an essential part of chess. Chess can be played without ratings.

Originally posted by ItalyBoyBlue


Ooooooh and nice comment about Italy, racist. When's the last time Scribblesland conquered the known world and more or less have dominated the artworld for most of the last 2000+ years?

Never, but I've also never teamed up with Nazi Germany in a world war or carried out atrocious inquisitions and crusades in the name of God.

Originally posted by ItalyBoyBlue
I'm afraid you're the one suffering from a severe misconception. And ELO ratings are by definition not an essential part of chess. Chess can be played without ratings.

Oh? Then why don't you just ignore them altogether rather than rallying for this silly statistic?

Originally posted by DoctorScribbles
Oh? Then why don't you just ignore them altogether rather than rallying for this silly statistic?

You're rediculous, and your unwavering fascination with my comments and this string are awe inspiringly pathetic...

Originally posted by ItalyBoyBlue
You're rediculous, and your unwavering fascination with my comments and this string are awe inspiringly pathetic...

What American university offers "upper division" math courses?

Originally posted by DoctorScribbles
What American university offers "upper division" math courses?

Every American University I know of offers them. Any class past 4th semester calculus is considered an upper division math class. But ELO is a stastically derived formula. IT'S ALGEBRAIC!!! So none of that matters does it. It's obvious you're only commenting to be obnoxious, rather than to provide insight into the issue. You've managed to completely muddy up this string but I will start a similar one in a while. I've sent a few messages to the administrators concerning your comments. I suggest you stop before they suspend your account.

Originally posted by ItalyBoyBlue
I suggest you stop before they suspend your account.

OK, I take it all back. Your idea is a good one. I hope it gets implemented soon.

As I am an expert in upper-division string theory I declare this string pointless.