Rating newcomers

Hi.

I'm new to Chess.com, and a friend of mine who is already rated, kindly offered to give me my first online game. It was a great game, very close, and I managed to win. My rating shot up from 1200 to 1606. His crashed from 1585 to 1505.

I've read the explanation of the rating system at

http://www.chess.com/article/view/chess-ratings---how-they-work

and Glickman's more coplete explanation at

http://math.bu.edu/people/mg/glicko/glicko.doc/glicko.html

The formula falls down, however, when a new person enters the system. Glickman suggests an entry point of 1500. Chess.com has chosen 1200, but the entry point is really irrelevant, because the rating for the incoming player is unknown.

There are three possibilities:

1 The new players real rating is very high, eg grandmaster, and the established player has a high likelihood of defeat

2 The new player is a beginner, and the established player has a high likelihood of winning.

3 The new player is somewhere in the middle.

In all three cases, however, the established player will lose or gain a great deal on one game if he is significantly above 1200 in rating and plays a new player. I feel badly that my friend lost 80 rating points agreeing to play me when (I think) we are very close in ability and I just happened to win the first game.

I realise that my rating will oscillate until it reaches a stable level. But I think there could be an adjustment to the formula so that the ratings of an established player are affected only minimally when he / she plays an unrated player. This could be done simply by reserving the rating of 1200 to new members only and reprogramming the formula to recognise a new player. (It does this, in part, already in assigning a maximum RD of 350 to an incoming player)

Otherwise established members are going to steer clear of playing new members until they have been rated.

I did enjoy the game, though.

And I was lucky! 

And thanks for chess.com. Its a great site.

i want to agree with my friend, but i think that it is a chance , a risk for the established player,  to end up playing a higher rated player. Again my friend is right, the established player should be protected against playing some one that is a grand master.  It is worth looking into, to improve the system for greeters.

Chess.com  is a great site.

Well Genghis, I don't feel bad that my rating went down and yours went up. You weren't THAT lucky!

In a way it's fair in this case that the winner has a better rating than the loser. (Such abstract rationalizations make me feel better.)

Good game nonetheless, and enjoying the rematch so far.

You really don't need to worry. It is the difference in rating that is more important then actual number. Ratings here on Chess.com are higher than in chess federations. Your friend will regain his points and since you said you are both about equally good. I don't see a problem.

PS

Glickman equation (for calculating winning expectancy) only takes difference into account.

(Of course you already know this, I just posted this simple explanation for people who usually doesn't take time to study "SEVERE NERD CONTENT")