|
|
|
Topic: alternative to Opponent Average Rating
| |
|
|
Author
| Message |
|---|
|
|
One thing that bothers me is that the average opponent rating doesn't mean much by itself, people will compare these numbers between players and that really doesn't tell you anything unless you pair it with the player's own rating.
I'm thinking that what this number really tells you is if you play stronger or weaker players than you, correct? Then we can convert this information into something similar. This new metric will be 0 when the average opponent rating is equal to your own rating. It will be negative if the AOR is lower than your rating and positive if it is higher. I haven't thought out all the details, but we can make it be something like (AOR - rating) / 100, so you will get 1 point for each 100 points above or below your own rating.
To me this looks much more understandable than the plain AOR. What do you guys think? Better or worse than plain AOR?
Thanks,
Miguel
|
|
|
i think both data yield the same result in gauging the opponent's strenght. yes, it'll be alright.
therefore, if a player has AOR of -100 and his personal rating stands at 1900, then you may assume that his plain AOR is something like 1800.
this is ok... 
|
|
|
Yes, but the advantage is that now it is valid to compare the values assigned to players, as this is a measure of the risk each player takes when choosing opponents.
Any opinions on including provisional ratings or not in this calculation?
Thanks,
Miguel
|
|
|
Yo estoy de acuerdo con parte de lo que se dice, y pienso que es importante lo de la comunidad de amigos. Por otra parte me gusta aprender a jugar. Y relativizo la idea del rating, porque pienso que por complicada que sea la formula para calcularlo, este debe ser orientativo, y no determinante (100 puntos arriba, 100 abajo, no determinan la calidad de un jugador.)
I es que ultimamente no hago ni tablas ...  , pero me gusta esta web, Miguel.
I agree with part of what is said, and I think that it is important that of the community of friends. On the other hand I like to learn how to play. And I "relativize"(is that the word?) the idea of the rating, because I think that for complicated that is the it formulates to calculate it, this it should be orientative, and not decisive (100 points up, 100 below, they don't determine the quality of a player. )
is that lately I don't make neither drawns. . . ; -), but I like this web, Miguel.
|
|
|
sir miguel, i think there must be something wrong with your formula in calculating provisional rating. i'm a bit apprehensive that it's over-inflated. if this cannot be remedied, i suggest there must be two AOR statistics separating provisional and permanent AORs.
best regards...
|
|
|
I brought this issue long time ago and had to agree that ratings changes should be calculated with the ratings at the begining of the game and not at the end. After consulting with the USCF I found out that I was wrong and ratings are calculated with the numbers AT the end of the game...
Similarly, I would suggest calculating the AOR with the current ratings irrespective of the fact that some of the ratings are provisional. Since AOR changes as your opponent's ratings change having one or two provisional ratings in the mix will not really affect the metric substantially as with time those provisional ratings will cease to be provisional (unless of course, some people start playing and leave a provisional rating in the database...)
Since Miguel is suggesting a percentage formula I believe that:
AOR quotient = [(Player's Rating-AOR)/Player's Rating]*100 (%)
If I understood Mandaragit he suggests one AOR without provisional ratings and one with them... I believe that one including all ratings is enough and wouldn't overcomplicate matters...
|
| 1 2 Next |
|
|
|
