Default rating of 1200 for new members

New members to the community receive a default rating of 1200. I assume it is because that is the expected rating of the average player. If you look at the frequency distribution of ratings, you will see that the curve peaks at 1200 and that it is skewed to the right. The fact that the curve is skewed to the right, indicates that the rating given to new members should be increased somewhat until the curve is symmetrical.

I think it is a very good comment, and I second that.

It must be said, though, that the rating system takes into account the fact that that is just a provisional rating, and variations in the first games are huge (you easily get +/- 100 points with one game): so, if you are stronger than that, it won’t take long to reach your “real rating”.

I believe the system is also taking into account the fact that if you have an established rating, and you play against somebody with no established rating, your rating won’t have a wild variation: this is to avoid that you lose too many points by playing against somebody who is actually much stronger than 1200, but given he/she just entered, his provisional rating is 1200.

nm

I've not seen the distribution of ratings, but would be interested to see it. If the curve peaks at 1200, then this is the mode of the distribution.  And a skew to the right suggests that the mean is somewhere above this.  My guess is that if all the players were reasonably active then the curve would become normal (shaped like a bell) and the skew would disappear. The shape may be due to the fact that some players start with good intentions, play a few games and lose, realise they've got a lot to learn and then lose interest in chess. This makes the distribution difficult to interpret and I'm not sure raising the default would affect this.

Hallo Jimbobins,

Jimbobins wrote: "I've not seen the distribution of ratings, but would be interested to see it."

To see the distribution of ratings, click on "My Home" and then on "Online Chess". On the menu at the top right of the page, click on the "View Players" item.

Jimbobins wrote: "If the curve peaks at 1200, then this is the mode of the distribution.  And a skew to the right suggests that the mean is somewhere above this."

Agreed.

Jimbobins wrote: "My guess is that if all the players were reasonably active then the curve would become normal (shaped like a bell) and the skew would disappear."

Correct.

Jimbobins wrote: "The shape may be due to the fact that some players start with good intentions, play a few games and lose, realise they've got a lot to learn and then lose interest in chess."

That sounds quite reasonable.

Jimbobins wrote: "This makes the distribution difficult to interpret and I'm not sure raising the default would affect this."

I do not agree with this. The curve peaks at 1200, because:

1. that is the default rating given to new members, and
2. many new mambers only play a few games and some probably do not play even one so that their ratings remain at or close to 1200.

If:

1. the default rating is increased to the actual average of the current members,
2. the behaviour of new members does not change, and
3. the number of new members with the new default rating exceed the number with the old default rating

Then the curve should become symmetrical.

Hi Andrea,

You're right.

Is it possible that the error in calculating the ratings would be less if one were working with a value closer to the actual average?

People who are rated well below 1200 and want a fresh start may decide to start a new profile, that could put points into the system too.

The thing to do would be to remove the users who have been inactive for a very long time, and slightly lower the rating of all users whose rating is still provisional so that the average is once again 1200.

I would also like to see the distribution of ratings. And I think it would be more useful if the new player ratings reflected a rating very close to the median for each style of play. This doesn't appear to be the case in Live Chess.

My concern about the 1200 rating for new players is that this number seems quite high for Live Chess games. 1200 may be a fitting median rating in Online Chess, I wouldn't doubt. But my own Live Chess rating for "Standard" length games is almost 1200 now (I'm at 1189) and this lands me at the 63.9 percentile. So if I'm ranking above almost two-thirds of players at this length of live games, I'm guessing that the median is actually around 1000 or 1050.

For "Blitz" length games, I'm now rated 1136, but that puts me at the 65.5% on here. So the median is clearly far below 1136. I'm guessing it's below 1000.

For "Bullet" length games, I'm now rated 1042 and I'm at the 46.4 percentile. So, I'm a little below average at these, and I'd guess the median here is almost 1100.

It seems to me that bullet ratings should begin at 1100, blitz at 1000, live standard at 1050, and Online Chess around 1200. But these are all guesstimates based on my own percentile standings and ratings for each level.

As it is, players who have never played on here get rated well above average to start in Live Chess. So when I beat them in Live Chess Standard or Blitz, I gain more points than against an average player. When I lose to them, I lose more points than I would on average.

When you lose to a lower ranked player, you drop hard and fast.

A FIDE rating has no relevance. A starting point point of 1,001,200 is the same as 1200 since only the chess.com rating matters to the pool of players in chess.com. The average player would then be 1,001,339 rather than 1339.

The starting point could even be zero with players currently below 1200 becoming players with a negative rating.

Iam sanjay kumar s  fide no 35002015  what is my rating  pls help me

lease be relevant, helpful & nice!

I do not believe the bell curve is relevant here. Bell curve is about normal distribution. But as there are fewer good players (above average) then not so good players (below average) I think one needs to look at an alternative probability definition in this case.

1189