Join Bridge Winners
All comments by Emily Middleton
1 2 3 4 5 6
You are ignoring the author of this comment. Click to temporarily show the comment.
Ok: (slightly different numbers as I've added the latest files and removed goulash)

4-4:
Both made: 1208
Both off: 421
4M made: 496
3N made: 492

5-3:
Both made: 2492
Both off: 660
4M made: 592
3N made: 868

All trump distributions:

Both made: 9509
5-3: 2492
4-4: 1208
6-2: 1059
5-2: 928
4-3: 791
5-4: 543
6-1: 486
6-3: 421
< 6 trumps: 312
4-2: 218
3-3: 156
7-1: 137
6-4: 106
7-2: 101
5-1: 84
5-5: 70
7-3: 60
6-0: 23
7-0: 22
6-5: 13
8-1: 11
7-4: 6
8-2: 5
8-3: 1

Both off: 3516
5-3: 660
4-3: 451
5-2: 429
4-4: 421
6-2: 317
5-4: 260
6-1: 199
6-3: 164
< 6 trumps: 83
7-2: 68
4-2: 64
7-1: 62
6-4: 44
3-3: 43
5-1: 43
5-5: 37
6-0: 30
7-0: 26
7-3: 14
8-1: 11
8-0: 4
6-5: 3
8-2: 1

4M made, 3NT off: 2810
5-3: 592
4-4: 396
6-2: 281
4-3: 254
5-4: 238
5-2: 228
6-3: 144
6-1: 130
< 6 trumps: 108
4-2: 65
3-3: 63
7-1: 44
7-2: 44
5-1: 36
6-4: 34
5-5: 33
7-3: 23
6-0: 10
7-0: 8
7-4: 4
6-5: 3
8-1: 2
8-2: 2
8-0: 2
7-5: 1

3NT made, 4M off: 4505
5-3: 868
4-3: 562
5-2: 542
4-4: 492
5-4: 307
6-2: 294
6-1: 241
6-3: 220
< 6 trumps: 194
4-2: 139
5-1: 92
3-3: 81
7-2: 79
7-1: 76
6-4: 59
6-0: 36
7-3: 28
7-0: 25
5-5: 23
8-1: 18
8-0: 8
6-5: 6
8-2: 3
7-4: 3
9-0: 1
8-3: 1
Nov. 17
You are ignoring the author of this comment. Click to temporarily show the comment.
I have a life, I'll be spending ~12 hours this weekend playing in-person bridge!

Wait, you didn't mean that did you ;-)
Nov. 14
You are ignoring the author of this comment. Click to temporarily show the comment.
There are also a few Zia goulash events (science vs nature and individual) and a youth goulash (bet that was fun!)

Without these, and a few other bogus ones such as the 12-1-1-1, we have the following:

10-3-0-0: 8
10-1-1-1: 10
10-2-1-0: 25
9-4-0-0: 37
8-5-0-0: 77
7-6-0-0: 137
9-2-2-0: 224
9-3-1-0: 272
9-2-1-1: 394
8-4-1-0: 1062
6-6-1-0: 1743
8-3-2-0: 2566
7-5-1-0: 2613
8-3-1-1: 2689
8-2-2-1: 4665
7-3-3-0: 6235
7-4-2-0: 8695
7-4-1-1: 9570
7-2-2-2: 12118
6-5-2-0: 15176
6-5-1-1: 16965
5-5-3-0: 21100
5-4-4-0: 29462
6-4-3-0: 31274
7-3-2-1: 44252
4-4-4-1: 70445
5-5-2-1: 75179
6-3-3-1: 81217
6-4-2-1: 111026
6-3-2-2: 133339
4-3-3-3: 247026
5-4-2-2: 250197
5-4-3-1: 304646
5-3-3-2: 365661
4-4-3-2: 507207

And some examples for the uncommon ones:

10-3-0-0: http://www.bridgebase.com/tools/handviewer.html?linurl=http://www.bridgebase.com/tools/vugraph_linfetch.php?id=13243 (Board 32, W)
10-1-1-1: http://www.bridgebase.com/tools/handviewer.html?linurl=http://www.bridgebase.com/tools/vugraph_linfetch.php?id=10464 (Board 18, W)
10-2-1-0: http://www.bridgebase.com/tools/handviewer.html?linurl=http://www.bridgebase.com/tools/vugraph_linfetch.php?id=1015 (Board 14, N)
9-4-0-0: http://www.bridgebase.com/tools/handviewer.html?linurl=http://www.bridgebase.com/tools/vugraph_linfetch.php?id=1110 (Board 8, E)
8-5-0-0: http://www.bridgebase.com/tools/handviewer.html?linurl=http://www.bridgebase.com/tools/vugraph_linfetch.php?id=12000 (Board 15, N)
7-6-0-0: http://www.bridgebase.com/tools/handviewer.html?linurl=http://www.bridgebase.com/tools/vugraph_linfetch.php?id=10168 (Board 25, S)
9-2-2-0: http://www.bridgebase.com/tools/handviewer.html?linurl=http://www.bridgebase.com/tools/vugraph_linfetch.php?id=10019 (Board 6, W)
9-3-1-0: http://www.bridgebase.com/tools/handviewer.html?linurl=http://www.bridgebase.com/tools/vugraph_linfetch.php?id=1003 (Board 9, S)
9-2-1-1: http://www.bridgebase.com/tools/handviewer.html?linurl=http://www.bridgebase.com/tools/vugraph_linfetch.php?id=10058 (Board 9, S)
Nov. 14
You are ignoring the author of this comment. Click to temporarily show the comment.
This includes all jumps including limit raises and 2+ level opening bids.

In the first 8 boards yesterday, against the opponent that prompted this comment, we had the stop card out 6 times:
1-1-1N- STOP 3N
1-1- STOP 4-4N-5-5N-6- STOP 7
1N-2-2- STOP 4
STOP 2N…
1-STOP 2N…

Perhaps if you're playing a precision-style system with more relays there would be fewer jumps?
Nov. 13
You are ignoring the author of this comment. Click to temporarily show the comment.
There are 9,513 boards attributed to a player named “Gold”. Some of them may not be David!
Nov. 13
You are ignoring the author of this comment. Click to temporarily show the comment.
7,764 occurrences of one table in 3N and the other in 5m
3,529 times they both made
1,467 times they both went off
942 times 5m made and 3N went off
1,826 times 3N made and 5m went off

I forgot to filter out doubled contracts - will have a look later and see what difference that makes.
Nov. 13
You are ignoring the author of this comment. Click to temporarily show the comment.
20,347 occurrences of one table in 3N and the other in 4M.
9,508 times they both made
3,520 times they both went off
2,811 times 4M made and 3N went off
4,508 times 3N made and 4H went off

Alan, I'll have a look at 4-4 vs 5-3 later.
Nov. 13
You are ignoring the author of this comment. Click to temporarily show the comment.
For the contract breakdown, same level is 1-7. For the lead stats I grouped together partscore, game etc.
Nov. 13
You are ignoring the author of this comment. Click to temporarily show the comment.
It's currently in JSON but I can convert to CSV easily enough. Send me a private message with your email address and your preferred format and I'll share it with you.
Nov. 13
You are ignoring the author of this comment. Click to temporarily show the comment.
You are ignoring the author of this comment. Click to temporarily show the comment.
In my defence we were *really* behind in this match. Does that make the opening forgivable?
Nov. 8
You are ignoring the author of this comment. Click to temporarily show the comment.
“ I don't actually see how any science can be applicable. Surely science requires a tight definition of the question.”

Perhaps I should have started this post with “I have *attempted* to apply some science, purely because I enjoy this sort of thing rather than because I think it will give a definitive answer” ;-)
Nov. 8
You are ignoring the author of this comment. Click to temporarily show the comment.
As several have pointed out above, my algorithm is indeed too simplistic. One tweak I have made is that if you are in “equivalent” contracts that both make, e.g. 2S= vs. 3S=, this imp is attributed to play rather than bidding.
(This shifted 8 imps to play on the given sample)

However, this tweak only considers contracts equivalent if they are partscore vs. partscore or game vs game. Regarding Debbie's example, I think 1N + 2 vs 3N -1 can still be argued as bidding, not play. Defenders behave differently when trying to score 7 tricks than when trying to score 5, so it is not unusual to make more tricks in a lower contract?

I think to really hone this I would need to hook up the analysis to a double dummy solver. For example, last night we bid and make 6 despite opponents having an AK to cash. Against my theoretical opponents in 4S, this imp swing would count as a bidding triumph rather than play. But, If deep finesse tells me the limit of the hand is 11 tricks, the category could be shifted accordingly.

Does anyone know of a freely available double dummy solver that can be called programatically? I have used deep finesse but only adding hands manually via the API.

I will tinker further and report back (when I'm not working, parenting or playing real bridge that is…)
Nov. 8
You are ignoring the author of this comment. Click to temporarily show the comment.
My mistake, edited :-)

I would be interested to run this over different matches and see how the percentage varies at different levels. However I am finding the vugraph archive hard to navigate. Plus I have real work to do!
Nov. 8
You are ignoring the author of this comment. Click to temporarily show the comment.
I've applied some science.

I've written a program that from vugraph archives:

a) looks at the contacts
b) looks at the imps gained/lost
c) assigned imps to the “play” category if the contract was the same, “bidding” otherwise.

I've ran this over 8 matches from the recent premier league (England open trials)

- 60 boards played in the same contract
- 127 imps gained on these 60 boards
- 68 boards played in a different contract
- 459 imps gained on these 68 boards

So that's about 22% play and 78% bidding.

Disclaimer - I knocked this up in my lunch break so may not be perfect.
Happy to analyse other matches!
Nov. 8
Emily Middleton edited this comment Nov. 8
You are ignoring the author of this comment. Click to temporarily show the comment.
You should remember I am always impeccably polite at the table Alan. Though this was an online match so partner's cup was out of reach!
Nov. 7
You are ignoring the author of this comment. Click to temporarily show the comment.
“Why oh why did we double the only contract we are sure to beat???”
I like doubling contracts I'm sure to beat ;-)

Perhaps I'm not following this comment - are you saying that you do not think the double of 4 was penalties, and that it prompted parter to bid when he may have passed otherwise?

My thinking was exactly the opposite (X was penalties and hoping to dissuade partner from further action)
Nov. 7
You are ignoring the author of this comment. Click to temporarily show the comment.
If he's only 6-4, he's either got 3 spades or 1 heart. The latter looks unlikely on this auction, and if he's got 3 spades why hasn't he supported you? (I didn't specify that 1 showed 5, will edit)
Nov. 7
RSI
You are ignoring the author of this comment. Click to temporarily show the comment.
Apologies if you clicked on this article expecting it to be about some new biding convention (red suit interference?)
Nov. 1
RSI
You are ignoring the author of this comment. Click to temporarily show the comment.
Sometimes rings but I tend to take those off and fiddle with them. Nothing on my wrist!
Nov. 1
1 2 3 4 5 6
.

Bottom Home Top