Mixed Strategies at matchpoints

[mr1303]

[hv=d=n&v=e&s=sxh98xxxd10xxcaqxx]133|100|Scoring: MP[/hv]

 

1H (Pass) ?

 

EW are sound players, but not top class. East hesitated for about 30 seconds before passing.

 

[hv=d=n&v=e&s=sxh98xxxd10xxcaqxx]133|100|Scoring: MP[/hv]

 

1NT (Pass) ?

 

1NT was 15-17

 

On both of these boards, me and my partner followed some mixed strategies and gained good scores here. I will explain what happened later, but for now I'd like to concentrate the discussion on which non-obvious bids you might make, and the likelihood of you making them, and the merits or otherwise of making these bids.

[Jlall]

4H and 2D followed by 3S.

[pooltuna]

4H and 2D followed by 3S.

agree 4♥ is obv. on the second one how about 4♦ followed by 4♠

[awm]

There are several interesting options here. On the first hand, besides the normal 4♥:

 

(1) 4♣ fit jump (if played this way) might be a good option.

(2) A bergen constructive raise followed by a game bid has some advantages.

(3) A psych of 1♠ is certainly a possibility (and it works a lot better after 1♥-P than 1♥-X).

 

On the second hand, 2♦...3♠ is certainly the normal action. Alternatives:

 

(1) 2♦ and then psych a minor suit splinter.

(2) 2♦ and then 6♥ (give the opponents a lead problem).

(3) 2♦ and then 3♦ (showing diamonds to try to deflect a killing diamond lead).

 

All of these lead to worse "double-dummy" slam bidding, but might work out in practice by convincing the opponents to make a poor lead.

[cherdanno]

I think not bidding 4♥ on the first one is terrible. RHO hesitated, so maybe LHO can't act when he might want to, so we are already ahead of the field. Of course, I may just be missing out on what kind of genius bid might work here.

[gnasher]

I'm too stupid to do anything but 4♥ and transfer-then-splinter.

[655321]

Yes, sounds like one of those irregular verbs.

 

I follow mixed strategies, You bid erratically, He can't bid to save his life.

 

:D

[ritong]

if i desperately need a good board, i would consider 3 NT on the second one, hoping to hit a case where the number of tricks is the same in both denominations, or opponents mislead, or are shocked enough with the dummy to misdefend.

[OleBerg]

1) 4♣ (fitbid) is not really an alternative, but rather a mistake I could make. (So according to me, there is a 1/30 chance that I would do it. According to my partner, the chance would be 1/6.)

 

I do not consider bidding differently because RHO hesitated, to be a mixed strategy. I would consider it to be "bidding according to circumstances". Likewise, if one of my opponents has to pass (for example because of a penalty card), I would preempt more conservatively. Not a mixed strategy, just bidding according to circumstances.

 

2) 2♦ followed by 3♠, splinter. And here I actually play, that partners 3NT is to play. So with my regular partner, anything else would be totally silly.

 

Otherwise, I echo the considerations of ritong.

 

I have yet to encounter a situation, where I would find a mixed strategy useful. But it could be an interesting desk-analysis.

[hrothgar]

I have yet to encounter a situation, where I would find a mixed strategy useful.

Have you ever made a psych?

 

If so, care to describe it...

[OleBerg]

I have yet to encounter a situation, where I would find a mixed strategy useful.

Have you ever made a psych?

 

If so, care to describe it...

I haven't psyched in serious bridge in the last 10 years. I play an aggressive system and is an overbidder. Furthermore, my partner heeds S.J.Simons advice, that says that when you play with an overbidder, you should overbid yourself.*

 

So our system generates enough volatility by itself.

 

But even if I decided to bluff once in a while, I would not consider it a mixed strategy. I would bluff when I felt the circumstances were right. Some of these circumstances might be difficult to quantify, but they would still be there.

 

 

 

* For easy reference, S.J. Simons advice is; If you play with an overbidder, and start to underbid to compensate, the overbidder will simply overbid even more, just to compensate. If, on the other hand, you overbid to, he will start to moderate his bid, to compensate for your overbidding.

[mr1303]

Well, on the first board, I did indeed psyche 1S, which caused the following auction:

 

1H P 1S P

2C P 4H (5D)

5H

 

all pass

 

This went 2 off, but was a complete top since EW were cold for 4S, and those who sacrificed in 5H were doubled. I don't have the full hands, but East had a borderline 1S overcall and West had a strong 2-suited hand with spades and diamonds.

 

On the second hand, my partner did raise to 3NT (as suggested by Ritong). Whilst on best defence this would be 4 off, the actual result was that opponents led a 4th best from their ace, and I claimed all 13 tricks, which again was a top since everyone else (that I asked) had had a Blackwood auction and stopped in 5H with 2 aces missing. A punt of 6H might be successful, depending on whether or not the opening leader fancied leading his ace or not.

 

What do people think of the merits of this style of bridge? In particular the second hand, which is not really a psyche, but more of a spin of the roulette wheel, just with (hopefully) slightly more favourable odds.

[jdonn]

What do people think of the merits of this style of bridge? In particular the second hand, which is not really a psyche, but more of a spin of the roulette wheel, just with (hopefully) slightly more favourable odds.

Top and bottom bridge is not winning bridge, I feel very strongly. The best way I can think of to explain why is this.

 

If you are a better than average player (you expect to get over 50% by playing "normally") then you have more to lose than to gain on the average board by shooting for tops and risking bottoms.

 

If you are a worse than average player (you expect to get under 50% by playing "normally") you will (clearly, I hope we can agree?) gain more by learning how to play bridge better than by shooting for tops and risking bottoms.

 

One more thing. Every bridge player in the world, right up to the very best, make tons of mistakes. (If online vugraph has taught us anything...) So it stands to reason that the best way for bridge players to improve their scores is to reduce mistakes.

[mtvesuvius]

I'm boring, but I wouldn't try anything other than 4♥ on the first and transfer then 3♠ on the second. I really try not to play top and bottom bridge as much as I used to, and recently I have won several matches against excellent players by just playing solid and not taking many risks (of course this is IMPs, but still...). My matchpoint game is usually relatively solid, so taking top or bottom views for me isn't a very attractive option. Of course I don't always play well, and if you're in a big event looking for a win, then I would agree with your decisions.

[jdaming]

Seems like especially with the long hesitation on the first board that bidding a practical 4♥ will give you the most protection against the long pause. In most cases I think I would probably be calling the director if LHO (unpassed hand) bid over it.

 

Agree on 2

[Mbodell]

What do people think of the merits of this style of bridge? In particular the second hand, which is not really a psyche, but more of a spin of the roulette wheel, just with (hopefully) slightly more favourable odds.

Top and bottom bridge is not winning bridge, I feel very strongly. The best way I can think of to explain why is this.

It depends on how favorable you think the odds are. For instance, if you think a weak nt is just a better system than a strong nt but the field plays a strong nt then both your expected value and your variance of results will go up if you play a weak nt. If you are far better than everyone in the field, this is bad. If you are anything else, this is probably good (unless it is a qualifier or something where half advance where again lower EV with lower variance might be better).

 

So lots of people play things like strong club, or weak nt, or different cue bidding style or what not and end up with more of their fair share of tops and more then their fair share of bottoms.

 

If you are good at figuring out the odds and do have the advantage then you'd be foolish, IMO, not to take it. If you were good enough to know that you'd get 55-60% on some board "playing down the middle" and 2/3 of the time get 90-100% and 1/3 of the time get 0-10% by taking the risky move, I'd take the risky move. That doesn't mean I'd take the risky move if it was 51% success and 49% failure.

 

I agree that overall the most important skill for me and many people is eliminating "routine" and "dumb" mistakes.

[hrothgar]

What do people think of the merits of this style of bridge? In particular the second hand, which is not really a psyche, but more of a spin of the roulette wheel, just with (hopefully) slightly more favourable odds.

Top and bottom bridge is not winning bridge, I feel very strongly. The best way I can think of to explain why is this.

 

If you are a better than average player (you expect to get over 50% by playing "normally") then you have more to lose than to gain on the average board by shooting for tops and risking bottoms.

 

If you are a worse than average player (you expect to get under 50% by playing "normally") you will (clearly, I hope we can agree?) gain more by learning how to play bridge better than by shooting for tops and risking bottoms.

Depends on what you mean by "more to lose"

 

For example: You only get masterpoints if you place towards the top in an event.

 

If your goal is to win as many masterpoints as possible - and you're essentially indifferent if you place 5th or 55th - then adopting high variance methods is very much the way to go...

 

As I recall, Gerben or Wayne did an interesting study a few years back showing the trade off between variance and expected value. It showed pretty clearly that its completely reasonable to lower expected value in favor of increasing variance.

[jdonn]

If you are good at figuring out the odds and do have the advantage then you'd be foolish, IMO, not to take it.  If you were good enough to know that you'd get 55-60% on some board "playing down the middle" and 2/3 of the time get 90-100% and 1/3 of the time get 0-10% by taking the risky move, I'd take the risky move.  That doesn't mean I'd take the risky move if it was 51% success and 49% failure.

This is all very interesting yet has nothing to do with practice, such as with the actual hands where no one has much clue at all how likely the actions chosen are too 'work' (My own estimate would be something like 20% in the first case and 40% in the second case.)

 

If your goal is to win as many masterpoints as possible - and you're essentially indifferent if you place 5th or 55th - then adopting high variance methods is very much the way to go...

All I can say is I can think of no pairs that I have ever met that usually score below 50%, but some periodic minority of the time bust out for a 65% game and the win.

 

Like with my above comment, it is much harder and people are much worse at predicting the chances of actions like 1♠ on that first hand working than at improving at 'solid' bridge.

[awm]

All I can say is I can think of no pairs that I have ever met that usually score below 50%, but some periodic minority of the time bust out for a 65% game and the win.

I know some pairs who are more or less like this. What's more common is pairs who seem to average around a 50% game but frequently do either much better or much worse. The most extreme example I can think of (Waldron+Waldron) seem to score below 40% and above 60% with roughly equal and extremely high frequency (I recall a regional where they played pairs for three days and every single session was one of these two extremes).

 

Arguably Wildavsky and Doub seem a little bit like this; they obviously were very strong in winning the US team trials last year, but their first two days in the summer nationals (LM pairs, then an open pairs when they failed to Q) were atrociously bad.

[mr1303]

That generally you should try to improve your sensible bidding and card-play is of course a sensible idea, and I don't think anyone is disputing that. And I'll continue to practice solid improvement on the 99 other boards out of 100 that I play.

 

Say I'm playing these two boards against the pair whom I believe to be the best pair in the room. Therefore, my score on these boards is likely to be worse than the score that all of the other North/South pairs, since the best pair are more likely to produce a successful action. If I make the "normal" raise to 4H on the first, is it not likely to be an auction duplicated at every other table in the room? Therefore, I am potentially disadvantaging myself if I produce the normal action, and so a mixed strategy/tactical bid/psyche is more likely to be successful.

[gnasher]

Say I'm playing these two boards against the pair whom I believe to be the best pair in the room. Therefore, my score on these boards is likely to be worse than the score that all of the other North/South pairs, since the best pair are more likely to produce a successful action. If I make the "normal" raise to 4H on the first, is it not likely to be an auction duplicated at every other table in the room? Therefore, I am potentially disadvantaging myself if I produce the normal action, and so a mixed strategy/tactical bid/psyche is more likely to be successful.

It depends upon how much you think their advantage is under normal circumstances, and what you think your expectation is if you do something offbeat. If, for example, you think that your expectation against this pair is currently 35%, and that an abnormal action will lead to a top 40% of the time and a bottom the other 60%, you should take the abnormal action.

 

On the first hand you quoted, I don't think this is the case at all. RHO has given his partner UI by pausing, and that may well already have neutralised their original advantage.

 

The second deal is closer, because your strong opponents may make a better lead than the rest of the field. An alternative strategy, which carries fewer risks than 3NT, is to eschew a slam try and just sign off in 4♥. That might well induce a less successful lead than at other tables where your hand splintered.

[gnasher]

I think that this sort of tactic comes up more in long IMP matches, where you may already be down, or may know that you're the underdogs.

 

During my brief appearance in the Spingold this year, I held x AQx KQJxxxxx x, first at equal vulnerability. I opened 5♦, which I think is in general a poor bid, but I bid it because it has a higher variance than 1♦, and we were playing against a much better team. That was 11 out when partner made a normal pass with a good hand, and we missed a slam. Even though a loss was more likely than a gain, I think that it was clearly the right thing to do under the circumstances. In fact, I wish I'd taken more such actions.

[655321]

Suppose every pair in the room discovers this secret of success against strong pairs. The strong pair now is constantly facing opponents who bid singletons, randomly bid 3NT instead of looking for 6M, and generally deliberately take anti-percentage actions on every board (in addition to all their usual errors).

 

I would expect the strong pair's scores to improve dramatically (which admittedly is just another way of saying that I think this is a terrible strategy). If your theory is correct though, the strong pair would get worse scores than they used to.

[awm]

It wouldn't surprise me if the strong pair's score was actually hurt by this, although we have to be careful because a lot of people's ideas about how to "swing" are actually very low expectation.

 

But say we have a field with one pair that's much better than the others. Further, suppose everyone decides to preempt and overcall extremely aggressively against this pair. This means:

 

(1) The pair in question almost never gets to use superior methods they might have for uncontested auctions.

(2) The pair in question's judgment might be thrown off because their opponents' competitive actions are all extremely wide ranging.

(3) The pair in question will fairly frequently play in an abnormal contract, getting pushed to game when no one else is there, "wrong-siding" contracts because of a takeout double auction, or defending a doubled partscore when other people are in game; this tends to reduce the advantage from superior declarer play as they are often not in the "field spot" (of course sometimes they are handed a top because their contract scores better than the field spot).

 

The style opponents have adopted is not particularly sound and may hand out more good boards than bad. But a lot of the good pair's natural advantages have been nullified. I wouldn't be surprised to see their expected score go from 65% to 60% or something like this. Of course, if people behaved the same way against an average pair, the average pair may well have their score go up from 50% to 55% because of the unsound competitive actions too.

[jdonn]

These situations always start with "suppose", yet the supposing so often turns out to be wrong. Which of us haven't heard the stories about teams that swung on a board because they thought they were behind yet found out they lost the match on that board? Anecdotally taking these kinds of actions in the bidding works terribly, and it would be pretty impossible to convince me other than what I have been witnessing for a decade and a half.

 

Adam do you mean Scott and his dad? If so, have you considered other reasons they might score that way? I can think of one...