2.5.2 How Precise is my Rating?

The discussion in the last section showed that the above all the precision of rating depends on the number of games. But the text simplified the reality essentially:

• There are more than a couple of players. But closer investigation (See section 5.4 The Formal Description of the Procedure..) shows that the precision of the rating does almost not depend on the number of players.

• The players win strictly alternately. This assumption is too optimistic.

• There are only games between players of equal strength. Games between players of essentially different abilities gives almost no information. But because the pairing is done by the swiss system, most games give essential results.

• It has shown that (like all humans) players have good and bad days. Therefore the games on one day are not independent of each other. Effectively there are substantial less games. This item differs from player to player and is hard to estimate.

Neglecting the last item, there is a simple rule of thumb (See section 5.4 The Formal Description of the Procedure..) for the precision of the ratings as a function of the number of effective games (See section 2.6 Fame and Glory Fading away....):

Uncertainty	0	10	20	30	40
0	infinite	153	108	88	76
50	68	62	58	54	51
100	48	46	44	42	41
150	39	38	37	36	35
200	34	33	33	32	31
250	31	30	29	29	28
300	28	27	27	27	26
350	26	25	25	25	24
400	24	24	24	23	23
450	23	23	22	22	22
500	22	21	21	21	21
550	21	20	20	20	20

Rating uncertainty as a function of the number of effective games.