[shyams]

Thread for summary and general observations.

[Apollo81]

A quick review of these threads seems to indicate that pass > double > 2♥, and that pass and double are both a lot better than 2♥.

[Fluffy]

there are summaries on the board 30, probably they belong in here, so I'll post mine here as well:


I am not sure I added them correctly but my numbers say:

A-B -24
A-C 97
B-C 125


2♥ a big big loser, and double is clearly better than pass, however I suspect I was biased towards punushing 2♥ at times, and that I made good decisions over the take out double based on knowing the full hand also.


Pass risked having to catch up in the bidding later, double risked some awful club contracts, but somehow they were dodged most of the time, but gained a lot by getting in quickly.

2♥ was a loser because it acomplished nothing, when we had 5-3 heart fit, we where shouted down by their spade fit anyway. While when we had a bigger one, nobody was stopping us finding it after the double.

3♥ won several small swings by preempting their bidding though, at MPs it was not nearly as bad, didn't make the count, but maybe even better than pass.

[helene_t]

The main reason for not doubling was (at least for me) that partner might go wild with clubs. But it doesn't really seem to be the issue with many (any?) of these hands.

[helene_t]

I made a spreadsheet that computes the sum of the votes. Anyone can edit so if you have a few minutes when your boss isn't looking over your shoulders you can type in some results. I have typed in the first two boards.

https://spreadsheets.google.com/ccc?key=0At...ZsbVFQc1E&hl=en

[Fluffy]

helene_t, on Jan 21 2010, 04:38 PM, said:

The main reason for not doubling was (at least for me) that partner might go wild with clubs. But it doesn't really seem to be the issue with many (any?) of these hands.

once I reached 4♣ where the clubs were 5-1 and 1100 were possible, but they would bid 4♠ before the other could double.

[Codo]

Before the sim I had the impression that pass was the worst bid and 2 ♥ is less evil then the majority thought (but still not good).

But if I got my numbers right:

AB: + 11
AC: +10
BC: -3

So pass is better then bidding and double is as bad as bidding 2 Heart.
Now this is surprising, because 2 ♥ should be worse. But I have the impression that in too many hands E/W with a spade fit will "save" N/S way too often, because they do not know about the defensive values in partners hand. And it is often said here that 4 ♥ is a kind of transfer to 4♠. I guess this is espacially true in white/red with two established fits.

So, if I compare my result to Adams, his E/W seldom bids spades over hearts when it was wrong, while in my analysis, they always do. Of course his "perfect" opponents make passing a much better choice then bidding, because we will never drive them to the wrong level. I am not so perfect, so I thought what I had done and I had made it wrong as E/W more then once.

But still: Passing was the winner- a real surprise to me.

[shyams]

Whew! I have just completed all 30.

My tabulation (non 2/1 style bidding)

A - B: (Pass vs. Dbl) wins +26 IMPs
A - C: (Pass vs. 2H ) wins +38 IMPs
B - C: ( Dbl vs. 2H ) wins +18 IMPs

Notes - based on my results:
- 15 boards were flat or nearly flat (i.e. all 3 comparisons yielded 0, 1 or 2 IMPs either way)
- Pass vs Dbl had 3 bid swings (big defined as >10 IMPs). Two went in favour of pass, 1 went for the double
- Pass vs 2H had 6 big swings!
- Dbl vs 2H had 4 big swings (and 1 more of 9 IMPs)
- I was unsure of leaving the double of 2♥ in #1. That's why I posted the "Borrowed hand" thread -- http://forums.bridge...showtopic=36781
   My thoughts were that the spade suit is good enough to play opposite even a singleton; but perhaps I was biased.
- I found it interesting that whenever East declares in spades, a ♠A opening lead by South is often good, sometimes mandatory and rarely bad. I need to use this in real play, whenever appropriate

[jdonn]

I have a net of:

a-b: -18
a-c: +55
b-c: +73

Making double the winner. I really don't care about 2♥ so I'll look at why the swings of 4 imps or more I have between pass and double occured. There are ten of them.

1. Pass won by 4 imps. Double let EW sign off in a minor lower.
4. Pass won by 11 imps. Double went for a number when clubs broke 5-1.
6. Double won by 9 imps. Pass went for a number since NS had to guess whether to compete later and hearts broke 5-0.
8. Double won by 12 imps. Double let EW bid a minor lower but they went for a number.
10. Pass won by 8 imps. Double got NS too high with their 11 card heart fit.
14. Pass won by 4 imps. Double let EW sign off in a minor lower.
18. Double won by 5 imps. Pass kept us out of the auction when we either make something or push them higher.
21. Double won by 12 imps. Pass kept us from finding our game.
27. Double won by 6 imps. Double let us push EW too high.
29. Double won by 5 imps. Double let us push EW too high.

So here are my generalizations of the 10 hands.
Of double's 6 wins, 4 were when after passing we don't bid enough. 1 was when we got our hand off our chest but had to guess later after passing. 1 was when EW got in more easily over the double but got in trouble.
Of pass's 4 wins, 2 were when EW got in more easily, and 2 were when we got too high. I'll also note pass won a number of 1 or 2 imp swings because double allowed EW to play the hand better.

So here is what I see based on that, and my final summary.
3 times double let EW into the bidding more easily, which lost for us twice and won once.
7 times double led to us either getting into the auction or bidding more than we would have, which won for us five times and lost for us twice.
Several times double helped EW play the hand better though it usually just cost 1 imp.

I find double to be the winner for the reasons I expected. When we have values it's important to get into the auction when possible, but double is way more flexible and safer than 2♥. This is also why I agree with doubling on moderate balanced hands very often.

My final note, this entire method of analyzing is completely inaccurate! I had a swing because EW played transfers over doubles, I had a swing because EW played bart, and there were quite a lot of hands where some people would respond to the opening bid and some wouldn't. So the moral, sorry Mike, looking at hands like this just doesn't work.

[helene_t]

Thanks AWM, Codo and Mohitz for entering your scores in the spreadsheet.

So far, pass seems to be the winner:

A-B: 58
A-C: 119
B-C: 56

[eyhung]

I find it ironic that the best thing to do against the majority of BBO forums posters is to let them get into trouble using their own constructive methods.

[Mbodell]

So going through all the hands with Jack trying not to be influenced by seeing the other hands does give results quite different than what many posters guessed at when they could see all the hands. The net result is not that similar in that X is the best, but it is quite close and X > 2♥ > Pass. Counting as a "swing" anything that was 6 IMPs or greater when looking in summary we see:

a-b was -10 over all 30 hands. The swings for b were 4, 8, 27, 29. The swings for a were 2, 13, 19, 21. Comparing a to b in matchpoints b gets 32/60 and a gets 28/60.

a-c was -2 over all 30 hands. The swings for c were 4, 15, 22, 27, 29. The swings for a were 13, 14, 19, 20, 23. Comparing a to c in matchpoints c gets 33/60 and a gets 27/60.

b-c was 9 over all 30 hands. The swings for b were 8, 19, 20, 23. The swings for c were 21, 22, 29. Comparing b to c in matchpoints b gets 30/60 and c gets 30/60.

The other thing I felt was interesting about this set is that the hands didn't really seem a true random sample of all possible distributions. So I'm not sure how confident I'd be on the conclusions. I'd say we should run another 30 hands but this time switch the N and W cards. On this board set the N hands have more HCP but the W hands have more shape by a pretty significant margin over 30 boards. The summary is as follows:

Averages for W

spades 2.633
hearts 2.767
diamonds 3.3
clubs 4.3
HCP 6.667
freakness 3.233

Averages for N

spades 2.833
hearts 3.167
diamonds 3.1
clubs 3.9
HCP 8.033
freakness 2.5

Also of note, only 10 of the 30 boards were flat across all 3 actions (in terms of IMP results).

[awm]

Here are my observations about the hands:

(1) It makes quite a bit of difference what we assume that E/W methods are.

(2) There are several hands where E/W have a game that's potentially difficult to bid, and we have to judge whether they'd bid it with/without competition. While arguments can be made that certain games are easier/harder to bid in certain sequences, I think this is often guesswork. However, it has a huge effect on the result since these are often big swings.

(3) There are many hands where west has a decision of whether to respond after 1♠-pass on a sub-minimum hand. I've taken the approach of having west bid "almost always" because this is my typical style at this vulnerability. This has lead to E/W "stealing" a N/S game after a south pass on several occasions, when game might have been reached after 1♠-Pass-Pass... but it also lead to E/W getting too high on a few hands after a south pass when east had extra values. Obviously other people may feel differently about these, but I do think it is important to be consistent (i.e. west always responds on flat 5-counts, or west never responds on flat 5-counts).

(4) There are several hands where south has to make the decision whether to pass or bid later, for example after 1♠-P-2♠-P-P. Given the colors, I do not think this decision is obvious, but it seemed to me that passing again was most consistent with the original decision to pass (i.e. a conservative strategy in competition). If south is going to bid, double should be better than 3♥ for the same reason that we all agree double is better than 2♥ in direct position...

(5) Mbodell's observation about the hands is interesting. It's possible this is just the result of a small sample size, but the differences are pretty substantial here and could easily effect the outcome.

[awm]

Here are the hands where my result differs substantially from Josh Donn's:

Board 2: West held ♠Q9 ♥J5 ♦A2 ♣KT98742. I evaluated this as an invitational hand, both jumping to 3♣ without competition and looking for game after 1♠-X. Josh just took simple preferences with this hand. While I agree that this hand is worse than another ten-count with seven clubs which comes up on board thirty (xx Kxx x AKxxxxx) is it really so much worse that it's not worth an invite (Josh considered the second hand a game force)?

Board 6: Josh had south balance with 3♥ after 1♠-P-2♠-P-P, and go for a number. I had south pass it out, which seems more consistent with the original pass. If south is going to balance it seems clear to balance with a double for the same reason double is better than 2♥ overcall in direct seat.

Board 15: Josh has south balancing again, but this time it works to N/S advantage. As I've commented before, consistency is more important than what exactly you do here, although I still think balancing with 3♥ specifically is very poor.

Board 18: Josh has south pass after 1♠-X-P-3♥-4♣? Maybe, but this seem doubtful to me; vulnerable at IMPs with a big fit usually people bid game? I had south raise to 4♥ and play there (one off).

Board 22: Josh's E/W play transfers over 1♠-X. This causes the eventual 3NT contract to be wrong-sided after the double, whereas I had the natural auction 1♠-X-2♣-P-2NT which right-sided the contract.

Board 25: I thought E/W would reach game, with east holding ♠K8432 ♥K ♦A2 ♣AKT93 opposite ♠QT9 ♥QJ9 ♦J83 ♣8742. Josh had E/W stop short, west bidding a forcing notrump in an uncontested auction and showing a "bad raise" in competition, both of which deterred east. Since the number of trick taken actually depends on whether south passes or doubles, this made a substantial difference in the scores.

Board 26: We disagree on the likely play and defense after 1♠-X-P-1NT-All Pass.

Board 27: We disagree on the likelihood of north doubling 4♠ after 1♠-X-2♠-X-4♠ on ♠T ♥9865 ♦A9752 ♣KJ4 (I thought north would pass, Josh thought double was clear). Also, Fluffy's point about the number of tricks in 4♠ may be relevant -- we both have spades making eight tricks, but Fluffy points out that on a fairly likely diamond lead (or heart lead for that matter) there are nine tricks. This impacts Josh's result more than mine, since I have a push board in 4♠ all around, whereas he has 4♠X in some auctions and 3♠ undoubled in others.

Board 29: Josh had east respond 1NT on his 4234 two-count, whereas I had him pass the opening. A matter of style obviously (and a 3♠ preemptive raise might also be possible).

[jdonn]

Yes obviously some of those were judgment. I'll make the following comments.

awm, on Jan 24 2010, 01:59 PM, said:

Board 2: West held ♠Q9 ♥J5 ♦A2 ♣KT98742. I evaluated this as an invitational hand, both jumping to 3♣ without competition and looking for game after 1♠-X. Josh just took simple preferences with this hand. While I agree that this hand is worse than another ten-count with seven clubs which comes up on board thirty (xx Kxx x AKxxxxx) is it really so much worse that it's not worth an invite (Josh considered the second hand a game force)?

It's not about how good the hand is. I think it's quite wrong to make an invitational jump response on a suit like that, as well as having Qx of partner's major which is also a huge downside. If partner had rebid his major I could raise it. It's funny that you note you miss game in partner's major after making an invitational jump, but still think it's right on a hand with a bad suit and such good support for his major.

Quote

Board 6: Josh had south balance with 3♥ after 1♠-P-2♠-P-P, and go for a number. I had south pass it out, which seems more consistent with the original pass. If south is going to balance it seems clear to balance with a double for the same reason double is better than 2♥ overcall in direct seat.

That's quite wrong for two reasons. Bidding hearts is better than on the first round since partner is much less likely to be short in hearts after spades are raised. And double is worse than on the first round since partner can't respond 1NT to play which I would expect to be a good result after an initial double.

Quote

Board 15: Josh has south balancing again, but this time it works to N/S advantage. As I've commented before, consistency is more important than what exactly you do here, although I still think balancing with 3♥ specifically is very poor.

Yes consistency is important but this just goes to my point that one big problem with passing is you often give yourself a guess of whether to act or not later when it's much more dangerous to act.

Quote

Board 18: Josh has south pass after 1♠-X-P-3♥-4♣? Maybe, but this seem doubtful to me; vulnerable at IMPs with a big fit usually people bid game? I had south raise to 4♥ and play there (one off).

I still think passing is right. South is a total minimum, the extra trump on a defensive hand with no singletons is unlikely to be the difference that makes a bad game into a good one.

Quote

Board 25: I thought E/W would reach game, with east holding ♠K8432 ♥K ♦A2 ♣AKT93 opposite ♠QT9 ♥QJ9 ♦J83 ♣8742. Josh had E/W stop short, west bidding a forcing notrump in an uncontested auction and showing a "bad raise" in competition, both of which deterred east. Since the number of trick taken actually depends on whether south passes or doubles, this made a substantial difference in the scores.

And aren't these actions quite clear? It's a 4333 6 count with no aces or king.

But like I said, the main thing I learned from doing this is there are too many variables to do this exercise accurately. I thought the most random decisions were to respond to the opening bid or not. I think there were several where I did and you didn't, and several where you did and I didn't. However there were still some lessons to me.
- I confirmed why double will work well when it works well.
- I confirmed why pass will work well when it works well.
- I confirmed that 2♥ is even worse than I thought it would be.

[awm]

jdonn said:

Quote

Board 2: West held ♠Q9 ♥J5 ♦A2 ♣KT98742. I evaluated this as an invitational hand, both jumping to 3♣ without competition and looking for game after 1♠-X. Josh just took simple preferences with this hand. While I agree that this hand is worse than another ten-count with seven clubs which comes up on board thirty (xx Kxx x AKxxxxx) is it really so much worse that it's not worth an invite (Josh considered the second hand a game force)?

It's not about how good the hand is. I think it's quite wrong to make an invitational jump response on a suit like that, as well as having Qx of partner's major which is also a huge downside. If partner had rebid his major I could raise it. It's funny that you note you miss game in partner's major after making an invitational jump, but still think it's right on a hand with a bad suit and such good support for his major.

Sure, we both missed game in the uncontested sequence here. But I don't like your auction much. You've shown neither the values nor the seven-card suit, generating the same sequence you'd have on a flat five-count. There are many hands where opener cannot move over 2♠ and game is icy; to give an extreme example ♠AKxxx ♥Axxx ♦x ♣Axx offers good play for a grand slam in clubs, yet would likely pass the 2♠ preference. In fact there are few (if any) hands which will reach game in your auction but pass the 3♣ invite, whereas there are many hands which would go to game only after 1♠-3♣ (and in most cases game is good). Of course there are hands like this one where game is missed in both cases too. Perhaps this is a win for "standard" bidding where we can have the uncontested sequence 1♠-2♣-2♥-2♠ showing a doubleton honor in spades, 4+♣, and 10-bad 12 points, but of course "standard" is highly non mainstream these days.

jdonn said:

Quote

Board 6: Josh had south balance with 3♥ after 1♠-P-2♠-P-P, and go for a number. I had south pass it out, which seems more consistent with the original pass. If south is going to balance it seems clear to balance with a double for the same reason double is better than 2♥ overcall in direct seat.

That's quite wrong for two reasons. Bidding hearts is better than on the first round since partner is much less likely to be short in hearts after spades are raised. And double is worse than on the first round since partner can't respond 1NT to play which I would expect to be a good result after an initial double.

We could redo this entire thing with the question being whether south should balance over 1♠-Pass-2♠. I do think that part of the issue here is how aggressively you back into the auction later. If you're likely to back in later, then acting immediately has more appeal (safer than backing in later), whereas if you're willing to simply pass this hand throughout (unless partner does something other than pass) then an initial pass looks a lot better.

jdonn said:

Quote

Board 25: I thought E/W would reach game, with east holding ♠K8432 ♥K ♦A2 ♣AKT93 opposite ♠QT9 ♥QJ9 ♦J83 ♣8742. Josh had E/W stop short, west bidding a forcing notrump in an uncontested auction and showing a "bad raise" in competition, both of which deterred east. Since the number of trick taken actually depends on whether south passes or doubles, this made a substantial difference in the scores.

And aren't these actions quite clear? It's a 4333 6 count with no aces or king.

Making the "bad raise" is quite clear if you play transfer advances here (which my E/W didn't). However, it's not clear to me that opener simply passes the bad raise. Give responder for example ♠QJx ♥xxx ♦Jxx ♣Qxxx and game seems really good (especially since south is unlikely to lead a heart for the tap holding the ace without the king). If the given hand is an absolute dead maximum for a bad raise I can certainly see passing, but I usually play the bad raise as more like 4-7 or a bad 8-count (i.e. less than "constructive raise") rather than utter trash. Similarly, whether to respond 1NT or 2♠ without competition is a matter of style, but I'm not convinced that after 1♠-1N-2♣-2♠-3♣ responder shouldn't bid game on the double fit. For example, give opener ♠KJxxx ♥x ♦Ax ♣AKQxx and game is looking pretty reasonable, and opener can't just bid game opposite a possible misfitting 5-7 count.

jdonn said:

But like I said, the main thing I learned from doing this is there are too many variables to do this exercise accurately.

Definitely agree with this point.

[jdonn]

First one: Yes standard wins when you get to bid your 7 card suit and you make a grand in it! West could also rebid 2NT on the second round and show his values that way, in fact I like that better in hindsight. That might even lead to east bidding 3♠ and reaching game. So at least you can't say I was biased.

I think I would gladly admit that in general standard bidding does better than 2/1 on invitational hands. It's not something that matters to me much at all, but obviously when an invitational hand comes up you will be able to point to it.

Second one: Are we discussing whether to balance, or what to balance if you do? I think I talked about one and you answered the other.

Third one: I didn't give much though to whether east would pass the bad raise or not since they would stop in 3 anyway. He very well may not pass.

[gwnn]

eyhung, on Jan 24 2010, 07:23 AM, said:

I find it ironic that the best thing to do against the majority of BBO forums posters is to let them get into trouble using their own constructive methods.

sorry I don't really understand what this means. if it means what I think it means (that many forums posters went for a numbers in hands/missed games/slams), then I think it's not ironic, but commendable, it means that they were not DD but SD.

[eyhung]

My comment was not regarding the analysis, but the fact that passing, giving E/W a free run, often led E/W into bigger trouble than doubling, according to the analysis of the three people listed. Maybe pass is better than double, but I certainly didn't like some of the constructive auctions given by those forum posters which led to pass being rated better.

[Jlall]

awm, on Jan 24 2010, 03:48 PM, said:

jdonn said:

But like I said, the main thing I learned from doing this is there are too many variables to do this exercise accurately.

Definitely agree with this point.

lol if only someone had said this before this futile excersize began... oh wait I did