600k boards from Vugraph: lies, damned lies, and statistics
(Page of 7)

Earlier in the week, this thread http://bridgewinners.com/article/view/bridge-is-sixty-percent-bidding/ prompted me to write some code to analyse the Vugraph archives. Because that's what I do for fun.

I'm not reinventing the wheel here, but I thought I'd share as it feels like I've got the start of something interesting. So, is there any appetite for statistics out there? What (if anything) do you want to know? (I've been pointed at a double dummy solver but am yet to hook it up)

Turn the page for some starters for ten!

P.S. please don't ask me what the point of this... there is none!

Milton Point Count

Let's start with the basics. What's the most points anyone has held on Vugraph? 32, in the 2016 Vanderbilt

North
AK3
AKQJ
AKJ4
AK

Match result was 5N+2 vs. 6N+1. Follow the deal at http://www.bridgebase.com/tools/handviewer.html?linurl=http://www.bridgebase.com/tools/vugraph_linfetch.php?id=43904 (board 22)

And the worst? This 7-high beauty from http://www.bridgebase.com/tools/handviewer.html?linurl=http://www.bridgebase.com/tools/vugraph_linfetch.php?id=33220 (Board 26)

West
72
6432
642
7643

The commentator asked: "has anyone ever seen a weaker hand that East's?" Not on Vugraph!

Full distribution of points across all boards - examples available on request:

• 0: 8711
• 1: 18296
• 2: 31652
• 3: 58371
• 4: 90815
• 5: 122893
• 6: 154720
• 7: 188572
• 8: 209886
• 9: 220266
• 10: 222577
• 11: 212350
• 12: 189403
• 13: 162582
• 14: 134363
• 15: 105001
• 16: 77512
• 17: 55482
• 18: 37916
• 19: 24631
• 20: 15331
• 21: 8819
• 22: 5095
• 23: 2488
• 24: 1274
• 25: 569
• 26: 283
• 27: 107
• 28: 40
• 29: 10
• 30: 9
• 31: 5
• 32: 2

Hand distribution

Looking for some wacky boards? One of the most distributional deals comes, surprise surprise, from a goulash tournament

West
Q
8
AKQJ10843
K106
North
A10952
QJ875432
East
AKQJ10965432
A9
South
KJ87643
7
97652
D

Full list (examples available on request):

• 12-1-0-0: 8
• 10-1-1-1: 15
• 11-1-1-0: 18
• 11-2-0-0: 18
• 10-3-0-0: 55
• 10-2-1-0: 74
• 13-0-0-0: 85
• 9-4-0-0: 116
• 8-5-0-0: 145
• 7-6-0-0: 254
• 9-2-2-0: 288
• 9-3-1-0: 369
• 9-2-1-1: 415
• 8-4-1-0: 1197
• 6-6-1-0: 1823
• 8-3-2-0: 2705
• 8-3-1-1: 2723
• 7-5-1-0: 2743
• 8-2-2-1: 4695
• 7-3-3-0: 6317
• 7-4-2-0: 8847
• 7-4-1-1: 9589
• 7-2-2-2: 12130
• 6-5-2-0: 15309
• 6-5-1-1: 16999
• 5-5-3-0: 21191
• 5-4-4-0: 29551
• 6-4-3-0: 31460
• 7-3-2-1: 44345
• 4-4-4-1: 70465
• 5-5-2-1: 75224
• 6-3-3-1: 81270
• 6-4-2-1: 111111
• 6-3-2-2: 133391
• 4-3-3-3: 247051
• 5-4-2-2: 250258
• 5-4-3-1: 304770
• 5-3-3-2: 365731
• 4-4-3-2: 507276

Disclaimer - some of these are bogus - e.g. the 13-0-0-0 deals are placeholders when the real deal was unknown

Per-player stats

You will always get someone at the local club claiming that they never get any good cards. But are some players really more blessed than others? Consider how I compare to some famous names:

• Pszczola - 10.07
• Helness - 10.06
• Forrester - 10.03
• Multon - 10.01
• Zimmermann - 10.01
• Helgemo - 9.97
• Garozzo - 9.97
• Brogeland - 9.97
• Mahmood - 9.96
• Drijver - 9.92
• Gold - 9.91
• Middleton - 9.76

So that's why I'm not a world champion - I just don't get the cards!

Disclaimer - names recorded by Vugraph are in no way unique or consistent, and is usually just a surname. I haven't made any attempt at normalisation. I probably haven't played enough on Vugraph to claim statistical significance.

Contract breakdown

As has been discussed extensively on the original thread, one can't tell from the raw statistics whether an imp gain was through play, bidding, or a mixture of both. But here they are anyway.

Same contract, same declarer, same lead: - 63,504 boards

• of which flat: 39,642
• total imps: 99,151
• imps/board: 1.56

Same contract, same declarer, different lead: - 82,125 boards

• of which flat: 43,038
• total imps: 169,331
• imps/board: 2.06

Same contract, different declarer: - 35,511 boards

• of which flat: 16,832
• total imps: 89,139
• imps/board: 2.51

Same strain, different level: - 85,515 boards

• of which flat: 9,997
• total imps: 587,664
• imps/board: 6.87

Different strain, same level: - 50,122 boards

• of which flat: 7,493
• total imps: 306,699
• imps/board: 6.12

Different strain and level: - 138,146 boards

• of which flat: 16,130
• total imps: 946,038
• imps/board: 6.85

Disclaimer - I have made the assumption that matches which results for an 'open' and a 'closed' room are imp matches. This is not necessarily the case.

I've seen others provide simulations about trump vs. non-trump leads. But what happens in real life?

Against partscore:

• Trump lead: 40,962 of which 15,435 went off (37.68% success rate)
• Non-trump lead: 252,339 of which 96,651 went off (38.3% success rate)

Against game:

• Trump lead: 34,519 of which 12,351 went off (35.78% success rate)
• Non-trump lead: 319,632 of which 117,591 went off (36.79% success rate)

Against small slam:

• Trump lead: 5,683 of which 1,794 went off (31.57% success rate)
• Non-trump lead: 47,765 of which 17,637 went off (36.92% success rate)

Against grand slam:

• Trump lead: 1,826 of which 528 went off (28.92% success rate)
• Non-trump lead: 5,375 of which 2,235 went off (41.58% success rate)

And finally...

Stoppage time

In my match today an opponent was complaining about stop cards and how many minutes of his life he's 'wasted' adhering to them.

There were 605,592 jump bids in 1,090,205 auctions - a total of about 10 weeks waiting for the stop card.