1

Let suppose an unrated player plays a FIDE rated swiss tournament where its opponent's ratings are e.g. 1300,2000,2100,2200,2300 and wins all of them, then his initial rating will be 2100 according to FIDE rating calculator and FIDE documentation confirms this:

If he scores more than 50%, then Ru = Ra + 20 for each half point scored over 50%

Here Ru is the (new) rating of the unrated player and Ra is the average rating of the rated opponents.

Clearly, this player performed much better than the calculated rating (any performance calculator admits that). And I believe many before me have seen such an issue and such a performance (or maybe slightly weaker than that) from an unrated player is not impossible since many just don't play rated games. I'm wondering what is the rationale behind this? Does it mean that FIDE wants players to play enough many games before they win any title? Or is there any other reason to not use performance for calculating the rating? Were there already some discussions about the initial rating calculation?

9
  • 1
    You'd have starting ratings of 3000+ happening often if you didn't cap it in some way
    – David
    Oct 24 '20 at 19:14
  • 2
    Any player who wins all the games should not have an initial rating but rather an estimator of a lower bound. Suppose the player somehow got into a GM tournament, and then swept the field against 2700 & 2800 rated players? Clearly, they would be far better than that 2100 initial rating he had. FIDE should wait until a player has at least 20 games with at least two wins and two losses before even trying to set an initial rating. Better would be 20 games with a score much closer to 50%. The other complicating factor is how consistent is the player. Does he beat very high rated players and lose t Oct 24 '20 at 19:15
  • @David, not really, first, I do not suggest to take performance directly, but I do not expect to use such a simple formula: +20 for any half-point above average! Second, performance 3000+ is quite unlikely to happen. For this problem, I expect to have seen such discussions earlier among decision-makers or at least some top players bring this issue. Also, the case I mentioned is just to bold the situation, imagine someone who is actually in master level but has to play with 1200-1800 players, since he/she doesn't have a rating, then ends up with 1500 after the end of the tournament. Oct 24 '20 at 23:57
  • @postasaguest, I agree with you in general and I'm wondering if this issue has never been discussed by well-known people in the community? Oct 25 '20 at 0:09
  • @SaeedAmiri why should a more complicated formula be more likely to be correct? At least in my opinion, simpler means better
    – David
    Oct 25 '20 at 23:04
2

I'm wondering what is the rationale behind this?

The rationale is very simple. 5 games is far too few to get a realistic and accurate assessment of a player's strength. That is why a new player's K factor remains at 40 until either they have completed 30 rated games. Additionally it remains at 40 for all players under the age of 18 with ratings under 2300.

Here is the relevant section from the FIDE Rating Regulations effective from 1 July 2017:

8.56 K is the development coefficient.
K = 40 for a player new to the rating list until he has completed events with at least 30 games.
K = 20 as long as a player's rating remains under 2400.
K = 10 once a player's published rating has reached 2400 and remains at that level subsequently, even if the rating drops below 2400.
K = 40 for all players until their 18th birthday, as long as their rating remains under 2300.
If the number of games (n) for a player on any list for a rating period multiplied by K (as defined above) exceeds 700, then K shall be the largest whole number such that K x n does not exceed 700.

A K factor of 40 allows for very rapid changes in rating. Hence the last clause above which seeks to limit the rating change in any one month to prevent see-sawing rating.

1
  • " 5 games is far too few to get a realistic and accurate assessment of a player's strength", Sure, that's true, however, we have an initial assessment anyways, then the question is why taking people that low? To avoid giving titles for a few games, they can add one more criteria, e.g. as suggested in comments they can say that a player should play at least 20 games. But completely ignoring the performance and considering average sounds way far from the realistic value of the initial rating. Oct 25 '20 at 0:04
2

I think FIDE doesn't want people to have excessively high initial ratings, as this may result in a player getting discouraged as they lose hundreds of rating points.

Performance ratings are particularly inaccurate when dealing someone wins all or most of their games, as compared to situations where someone is close to 50%. A performance calculator would take your scenario and give the guy a performance rating of maybe 2900-3100. But I don't think it's reasonable to look at those results and say that they're likely a world-champion level player.

I'll agree that in the scenario you proposed, a 2100 rating is far below what is deserved. If you drop the win against the 1300 player from that scenario and replace it with a loss to a 2400 player, the initial rating changes from 2100 to 2260 despite that performance being objectively worse. FIDE chose simplicity over accuracy in some ways; they use average opponent's rating without looking at the individual opponents when calculating initial ratings.

Also note that scoring 0/5 or getting an initial rating under 1000 will result in you simply not being rated. This has something of an inflationary effect on ratings, as a particularly bad first tournament may not be rated at all, while a particularly good first tournament will. Perhaps the limiting of how many points you can get from a good first tournament counteracts this.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.