I'm aware that the FIDE ratings list is only published once per month. I'm interested in how often the ratings are recalculated internally to FIDE without being published.
Are ratings adjusted match after match, or after every tournament? Even if the ratings are updated once a month, does FIDE count every match, or a whole tournament?
An example to make my point clear
Consider a three-player, single round robin tournament.
Roanin Novel 2798
Rigi Hains 2745
Dan H. Yvasin 2784
Round 1: R. Novel (1) beat Rigi Hains (0)
Round 2: Rigi Hains (0.5) drew with D.H. Yvasin (0.5)
Round 3: D.H. Yvasin (1) drew with R. Novel (1.5)
If FIDE was calculating the players' ratings, would it first update the players' ratings match by match1, or would their rating be adjusted by counting the tournament as a whole?2
1 In this case, rating changes after every match.
2 Here we calculate the expected score of each player at the start of the tournament, and compare it with the final score only. No rating changes occur between matches.
The calculations (all assume K=10)
Method 1 [Match by match] (Final ratings in bold)
After the first match, R. Novel is rated 2802.24 and Rigi is rated 2740.76.
After the second match, Rigi is rated 2741.38 and Dan is rated 2783.38
After the last match, Dan is rated 2783.65 and Novel is rated 2801.97
Method 2 [Whole tournament]
Initial expected score = 1.096
Actual score = 1.5
Hence, new rating = 2802.04
Initial expected score = 0.868
Actual score = 0.5
Hence, new rating = 2741.32
Dan H. Yvasin
Initial expected score = 1.036
Actual score = 1.0
Hence, new rating = 2783.64
In such a short tournament, there doesn't seem to be a significant difference in the results of the two methods, but maybe there could be one in a longer tournament.
Which of them does FIDE use?