The best practice for scoring a multisession multiple team competition
We have a multiple teams competition that is held over six sessions (once every two months) with the best four sessions determining the winners. For each individual session there are normally fewer than five boards played in each mini-match. The session is about seven tables. Currently each session is scored using IMPs in accordance with EBU guidelines for a single multiple teams session with four or less boards per match. Thus those with the highest aggregate IMP score over for their best four sessions wins the tournament and the right to represent the club at the County interclub teams competition.
However, over six sessions it is possible that one team wins one session with a large IMP score which coupled with moderate results in the other five sessions wins them the competition.
Is there any method of scoring that reduces large session IMP swings, that is fair to all participants and is simple to implement electronically through our scoring system (Scorebridge). The obvious one is using VPs for the mini-matches, but this falls foul of the EBU recommendation for low board matches.
Help and guidance appreciated
Comments
Both Victory Points and 'capping' would have the effect of limiting large wins. We say that Victory Points are unsuitable for <5-board matches, but capping is a appropriate. We have some formula for suitable caps, but you could look at the data from previous sessions and pick some multiple of 5IMPs which corresponds to a big win.
Have you experienced problems with your existing process? It seems to me that you will have about 36 matches of four boards in each season, so I would be surprised if any one of those matches had a disproportionate effect.
The other solution given in the White Book is to have a cap on the number of IMPS that can be won in a match, and perhaps that would be the answer?
There are a couple of teams that have dominated the competition with large wins in a session. So was simply looking for a fairer methodology
Capping looks the way to go
Could you please supply your formula for capping.
I will also look at your suggestion of looking back on the previous competition to decide on a "big" win
We need to make a decision shortly about how to score the competition going forward.
Quick answer: cap for matches of n boards 15sqrt( n ); for 4 boards this gives 30IMP.
This is the formula for IMP for 20.00-0.00 VP win in the WBF scale (which EBU uses for n > 4 ).
Long answer: EBU conditions of contest for trials has a formula for caps on pairs cross imps, but not for teams IMPs. The formula for pairs cross imps involves factors of sqrt(2 rc ) where r is the number of results, and c is the number of comparisons ( c = r - 1 ). Stripping out those factors from the cross imp formula gives 15sqrt( n ).
This formula is all well and good. But for very small n , some common sense is required. n = 3 gives 25, is OK but then (for smaller n ) formula gives capped total IMP less than 24, which feels wrong.
Thanks Robin
I will see how well capping or VPs goes down first of all with the committee and then see what the players think of it.