Opp-strength-calibration in large pairs tourneys

[helene_t]

Previous weekend Shogi and I managed to end in the prices (4th of 200ish) in a large pub crawl in the Netherlands.

 

However, I don't believe our results were the 4th best. Of the 7 pairs we opposed, 6 scored below average and 2 were extremely poor scoring barely above 30%. Such anormalities are inevideble in very large tournaments without any prior info that can be used for seeding.

 

Now it could be that we were compensated for this by having more strong pairs playing in our direction. But there is no particular reason to think that was the case.

 

Obviously, we should somehow be compensated for the strength of the opponents and for the strength of our direction. This is not trivial since you don't want to fit the noise. Ideally, someone should let a Bayesian network chew the data.

 

This is a bit similar to the national ranking system of the EBU, but with microdata.

 

Anyone who knows if this is ever done?

[Zelandakh]

How many boards did you play? Surely the solution is to play against a larger proportion of the field. Oh yes, and congratulations! The beers are on you! :D

[barmar]

If you faced lots of weak pairs, so did most of the other pairs sitting in your direction. If they gave you gifts, you can expect that they were also giving them out to others. So this evens out somewhat.

 

Remember, you're mainly competing against the people sitting in your same direction. The people who got a big boost from this uneven distribution of strength were the few strong pairs sitting in the other direction.

[helene_t]

If you faced lots of weak pairs, so did most of the other pairs sitting in your direction. If they gave you gifts, you can expect that they were also giving them out to others.

it's not a partial Mitchell it's a scheveningen movement which is similar to 1/30 Howell in this case. so few other pairs will have had as easy opposition as we.

[helene_t]

How many boards did you play? Surely the solution is to play against a larger proportion of the field. Oh yes, and congratulations! The beers are on you! :D

well no since more than 7 opps would be impractical. 4boards per rround is an absolute minimum if you don't want too much of the day spent on reading convention cards. this is not ebu culture :)

[Vampyr]

well no since more than 7 opps would be impractical. 4boards per rround is an absolute minimum if you don't want too much of the day spent on reading convention cards. this is not ebu culture :)

 

A pub bridge event is for fun, and anyway the prizes weren't vast sums, were they? You may be overthinking this. But it might have been nice for the event to be flighted, so that more people could get prizes.

[helene_t]

A pub bridge event is for fun, and anyway the prizes weren't vast sums, were they? You may be overthinking this. But it might have been nice for the event to be flighted, so that more people could get prizes.

Heh, they achieved this to give two bottles of wine to every pair whose rank was a multiple of five, lol :) This may sound silly but it is important to motivate a broad range of pairs to participate. So giving all the prices to the card sharks would not work.

[gordontd]

it's not a partial Mitchell it's a scheveningen movement which is similar to 1/30 Howell in this case. so few other pairs will have had as easy opposition as we.

I'm not quite sure what you mean about the movement - is it a partial everlasting Howell? If so, I doubt you could do anything to ensure much in the way of balance. Otherwise, you could seed starting positions so that each group is roughly as strong as the others - we do this for EBU pairs events where entries are known in advance.

[helene_t]

I'm not quite sure what you mean about the movement - is it a partial everlasting Howell? If so, I doubt you could do anything to ensure much in the way of balance. Otherwise, you could seed starting positions so that each group is roughly as strong as the others - we do this for EBU pairs events where entries are known in advance.

I think the movement is probably as good as it can be given the practical constraints (only meeting 7/200 opps, not much useful info about strength). Sometimes they do segregate on the basis of self-described skill level (like: kitchen players, club players, 2nd/3rd class league players, 1st/main, higher). In this case they didn't segregate.

[maartenxq]

When I used to play big dutch mps tournaments, I was always asked to indicate our pairs playing strength. This could and can be controlled by yr nbb masterpoints or competition level you play.

Based on this data it is quite easy to make a number of groups (like now in champions league) from which you draw the playing list.

 

I do not know however how 'sophistacated' yr pub drive operates.

 

These drives have of course a large luck factor. Years ago our rather mediocre team won the Amsterdam pub drive, among other factors because in our last match the weakest player of our oppposing team chose to sit west instead of east and make a mess of all the easy decisions he had to make.

 

Greetings Maarten Baltussen