[lamford]

BOARD 1                                                 
NS  EW  Contract        Lead    NS+     NS-     XIMP    XIMP
======  ==========      ====    =====   ========        ===
1   8   3S=     E       DA              -140    14      -14
2   4   3S+1    E       DA              -170    10      -10
3   7   4S-1    E       DA      50              24      -24
4   2   4S=     E       C2              -420    -10     10
5   6   3H-1    N       SA              -50     1       -1
6   5   4S-2    E       DK      100             -1      1
7   3   4S=     E       DA              -420    -24     24
8   1   4S=     E       DK              -420    -14     14

This is board 1 from one of the two section in this weekend's Tollemache Final:


I hope the tabbing is good enough, but the XIMP for 4S= is different in one match to another. Is the XIMP average just the average of these figures? If so, it seems meaningless, as you do not know whether one pair had a good result or the other had a bad result.

[Trinidad]

I think I know how X-IMPs work, but I can't make anything decent out of those figures. It seems clear that somebody must have chosen a wrong option somewhere in the scoring program, but I can't figure out what it is.

Rik

[lamford]

Trinidad, on 2015-February-15, 16:34, said:

I think I know how X-IMPs work, but I can't make anything decent out of those figures. It seems clear that somebody must have chosen a wrong option somewhere in the scoring program, but I can't figure out what it is.

Rik

Thanks Rik. That was my thought, and a later board where we lost 16 XIMPs for cashing two aces against a slam after opponents had a Blackwood mixup, also made no sense. Teammates also bid a non-making slam, and another team mate had a great result of -300, while the fourth pair were stuck in traffic and lost 4 IMPs on the board. Perhaps RMB1 can help. Or maybe the XIMPs are just that, the XIMPs in the match only, in which case they are meaningless.

[gordontd]

Where did you get those figures from? If you look at the results displayed online it's perfectly clear how they were calculated.

[gordontd]

Actually now I see that you have just displayed the traveller from one section. If you look at the match results displayed just below the ranking for each session, you will see the four results that make up the final crossIMPs score for each team. The way it works is that you score each board four times - AA, AB, BA, BB - and add the four scores before converting to VPs. The scores you showed won't add up because they are missing half the data.

[lamford]

gordontd, on 2015-February-15, 16:49, said:

Actually now I see that you have just displayed the traveller from one section. If you look at the match results displayed just below the ranking for each session, you will see the four results that make up the final crossIMPs score for each team. The way it works is that you score each board four times - AA, AB, BA, BB - and add the four scores before converting to VPs. The scores you showed won't add up because they are missing half the data.

Yes, those are shown correctly, and I am aware that the XIMPs for all boards are summed before converting to VPs. However, it seems to my untutored eye that the XIMPs for each pair are the average of their XIMPs per board for the event. Is this the case?

[barmar]

lamford, on 2015-February-15, 16:01, said:

If so, it seems meaningless, as you do not know whether one pair had a good result or the other had a bad result.

Isn't that true for all duplicate scoring? You can never tell if a good score is because that pair did something great or because their opponents gave them a gift.

[gordontd]

lamford, on 2015-February-15, 16:53, said:

Yes, those are shown correctly, and I am aware that the XIMPs for all boards are summed before converting to VPs. However, it seems to my untutored eye that the XIMPs for each pair are the average of their XIMPs per board for the event. Is this the case?

I think you are confusing two different things - the basic method of scoring is To8 scored by cross-IMPs, which is scored as already described. Then there are also Cross-IMP results for each player, where they are scored against every other table in the field and those scores averaged. This produces the Cross-IMP ranking list for all the players, which is just for interest and is not an official score of the event.

[lamford]

gordontd, on 2015-February-15, 17:01, said:

I think you are confusing two different things - the basic method of scoring is To8 scored by cross-IMPs, which is scored as already described. Then there are also Cross-IMP results for each player, where they are scored against every other table in the field and those scores averaged. This produces the Cross-IMP ranking list for all the players, which is just for interest and is not an official score of the event.

I understand, thanks. I find online that one XIMPs with every pair in the opposite direction and averages the result, as you say, but the site does not produce an individual record.

[PhilKing]

Maybe I'm being a bit thick, but the figures look like a total joke. (Mine look to be wrong by about an imp a board for the Saturday).

[lamford]

PhilKing, on 2015-February-15, 18:47, said:

Maybe I'm being a bit thick, but the figures look like a total joke.

I understand them now, and think they are what they say they are. We lost 56 XIMPs at our table in one 14 board match, when the opponents made five games rarely bid elsewhere and found a cheap sac!

[PhilKing]

lamford, on 2015-February-15, 19:10, said:

I understand them now, and think they are what they say they are. We lost 56 XIMPs at our table in one 14 board match, when the opponents made five games rarely bid elsewhere and found a cheap sac!

Well I have gone through all of Saturday and mine are wrong by a huge margin.

[nige1]

Imping against a mean or median score is widely reviled as being inferior to cross-imps. Such scoring methods, however, have the merit of simplicity. Averages are easy to check. Also, you can easily check your imp-score using the Bastille scale, in real time -- rather than discover scoring errors, after the correction period.

[gordontd]

PhilKing, on 2015-February-15, 19:16, said:

Well I have gone through all of Saturday and mine are wrong by a huge margin.

When I get to the office and have the files I'll send you a breakdown.

[lamford]

PhilKing, on 2015-February-15, 19:16, said:

Well I have gone through all of Saturday and mine are wrong by a huge margin.

You are right, and I think your correct XIMPs were +94.41, and you should have been second behind Tom Townsend and Nick Sandqvist on the XIMPs. The error could have been Tom's, to be fair to the organisers, in that he may have shown you in the wrong seats for the Middlesex match. Or you might have sat in the wrong seats ...

[PhilKing]

lamford, on 2015-February-16, 06:47, said:

You are right, and I think your correct XIMPs were +94.41, and you should have been second behind Tom Townsend and Nick Sandqvist on the XIMPs. The error could have been Tom's, to be fair to the organisers, in that he may have shown you in the wrong seats for the Middlesex match. Or you might have sat in the wrong seats ...

I sat in the right seat in that I played TE and W.

It's all sorted now.

[barmar]

nige1, on 2015-February-15, 20:48, said:

Imping against a mean or median score is widely reviled as being inferior to cross-imps. Such scoring methods, however, have the advantage of simplicity. Averages are easier to check. Also, you can easily calculate your imp-score using the Bastille scale, in real time -- rather than discover scoring errors, after the correction period.

Imping against a datum (Butler IMPs) was common in the days before computer scoring, because it required far less calculation. X-imp is quadratic, Butler is linear. But it was generally considered a compromise for practicality. Once computers took over, there was no need for it.