[hrothgar]

tysen2k, on Oct 7 2005, 01:19 AM, said:

Some good guesses. I'll let more people chime in before showing you the actual hands. But remember that there's another part to this. How are we going to take this data and create some rules/formulas for determining preempts?

Sadly, we aren't...

I think that the notion of "par" is interesting, however, its far too restrictive for system design. To raise one very simple objection, the structure that you have currently defined is based on an assumption of single suited preempts. If we started to include two preempts like Muiderberg or assumed fit methods like Ekrens we'd need to create a matrix describing par bids for a wide variety of different contracts. In turn, this substantially increases the complexity of the analysis.

Equally important the notion of par ignores issues related to the "speed" of the bidding. Its entirely possible that bashing to suboptimal contracts will score much better that a slow scientific approach that gets you much closer to par. Once again, consider "assumed fit" methods like Ekrens. If I open 2♥ a hand like

♠ AJ843
♥ KT52
♦ 872
♣ 9

its entirely possible that we'll miss par. I could play in a 2♥ on a 4-3 fit with a 5-4 fit available in Diamonds (or worse yet in Spades). However, this doesn't matter if the opening leads to my scoring +500 defending 3NX.

[awm]

Well some of this depends a bit on form of scoring (i.e. how bad is it to go beyond par and be doubled for a number -- at MPs it's bad but at IMPs it may not be too costly in some cases).

However, it seems reasonable to conclude:

(1) Hand three is a bad candidate for a (single-suited) preempt. It is quite likely that spades is not your best strain. While 2♠ is fairly likely to be "safe" (not beyond par) it's also fairly likely that you'd be better off in another suit. Of course, if the preempt gives more information than "I have a bunch of spades" then it might be okay.

(2) Hand one should almost surely be opened 3♠. It's very likely that you belong in spades, and 3♠ is not likely to be too high. On the other hand, opening 4♠ will already be past par something like 42% of the time, and is probably too aggressive. Hopefully partner can figure out to raise 3♠ to 4♠ on those hands where par is four spades or above.

(3) Hand two is a good candidate for a spade preempt, as again it is very likely that you belong in spades. A 4♠ opening is reasonable (this is par fully half the time and beyond par only 26% of the time). It may be "safer" to open 3♠ and let partner raise to four, but opening 4♠ is definitely fine.

Basically there's two things to determine:

A. Is this hand a good candidate for a preempt? Basically this means, is there a high probability that you belong in a particular suit (or suits in the case of two-suited preempts)?

B, How high is safe? In other words, how high can you bid such that the probability of exceeding par is small?

Of course, there's some question as to what a "high probability" actually means. My feeling is that A should be that you belong in your suit at least 80% of the time, and that B should mean you're past par at most 25% of the time... but that's just a rough estimate and will depend on things like the exact form of scoring and how aggressively opponents can manage to double (i.e. if you're past par only because of placement of opponents cards they may be unlikely to double you).

[cherdano]

The more I think about this, the more I am convinced that we can absolutely not derive a "preempt index" just from looking at the par of our side.
Imagine a hand whose par is 3 or 4 spades, but the opponents par is typically 2♦. There is nothing wrong with opening 1♠ and letting partner evaluate his hand to decide whether to bid game. Now imagine the opponents' par is 5-7 ♦. There is a good case for preempting with 4♠, as even if we are beyond par, opponents may not be able to double. And more importantly, they have little space to find out at which level they belong.

I am wondering whether a similar method to the one used by Tysen in his bidding system design contest might work. However, it would have to factor in at least two more things:
1. How likely opponents are to double rightly or wrongly (given their hand and our preempt definition). If preempts would always get doubled when that is right, we would obviously preempt a lot less.
2. Partner will jump to the most likely par from his side when it is his turn (we may assume for simplicity we cannot have a constructive auction after our preempt), and this should somehow get taken into account when determining opponents' available space.

Of course, this wouldn't give a preempt index for single hands, but for preempt definitions.

Arend

[mikestar]

I'm a bit confused about "our par contarct". My understanding is that there is one par result for both sides: that contract where neither side can better its score by bidding on or doubling.

For example, let's say that with neither vulnerable, NS can take 10 tricks in hearts and EW can take 10 tricks in spades. Now any partscore is not par, as etiher side can score more by bidding their game. 4♥ by NS is not par, EW can better their score by bidding 4♠ for a double game swing. But NS can better their score for letting 4♠ make by bidding 5♥. Then NS can better their score by doubling.

Par is 5♥X-1 by NS for NS -100.

Cleary EW can do no better: bidding 5 spades turns a plus into a minus. NS can do no better: a redouble or a higher bid turns a minus into a larger minus.

Par assumes perfect information (as a double dummy solver has). With imperfect information, it is possible to beat par. For example, NS might not take the push to 5♥ because they didn't know they have even 9 tricks in hearts, then EW beat par.

[Blofeld]

My understanding is that "our par contract" means:

(1) the par contract, if it is ours
(2) the highest contract played by our side such that we would prefer it to the par contract, if the par contract is them making something
(3) our best contract, if the par contract is them sacrificing.

But that's a rather contorted definition. So perhaps my understanding is out.

[tysen2k]

Blofeld, on Oct 7 2005, 04:11 PM, said:

My understanding is that "our par contract" means:

(1) the par contract, if it is ours
(2) the highest contract played by our side such that we would prefer it to the par contract, if the par contract is them making something
(3) our best contract, if the par contract is them sacrificing.

But that's a rather contorted definition. So perhaps my understanding is out.

That's it except for one small nit on point #2. Our par is the lowest possible sacrifice that's still profitable. Let's say we can take 11 tricks in hearts and they can take 11 tricks in spades. If you took your definition #2 literally, then it would say that our par is 7♥, when it should be 6♥.

[tysen2k]

tysen2k, on Oct 6 2005, 10:19 AM, said:

 [space] [space] Hand A [space] Hand B [space] Hand C
7x [space] [space]3% [space] [space] [space] 0% [space] [space] [space] 0%
6S [space] 10% [space] [space] [space] 7% [space] [space] [space] 3%
6x [space] [space]3% [space] [space] [space] 3% [space] [space] [space]11%
5S [space] 16% [space] [space] [space] 7% [space] [space] [space]11%
5x [space] [space]3% [space] [space] [space] 0% [space] [space] [space] 3%
4S [space] 23% [space] [space] [space]50% [space] [space] [space]14%
4x [space] [space]0% [space] [space] [space] 3% [space] [space] [space] 9%
3S [space] 26% [space] [space] [space]13% [space] [space] [space]26%
3x [space] [space]0% [space] [space] [space] 3% [space] [space] [space] 9%
2S [space] 16% [space] [space] [space]13% [space] [space] [space] 9%
2x [space] [space]0% [space] [space] [space] 0% [space] [space] [space] 3%
1S [space] [space]0% [space] [space] [space] 0% [space] [space] [space] 0%
1x [space] [space]0% [space] [space] [space] 0% [space] [space] [space] 3%

Repeating the table for convinience...

Okay, hand A was "the crapshoot in level, but obviously belongs in spades":
KQJxxxx (xxx xx x) --> (side suits in any order)

Not really any surprises there. But look at hand B which was much narrower range of levels and much more likely belonged in 4♠:
(KQJxxxx (Axx xx x)

The addition of a side ace had a big impact on the par level. Looking at the par data most people felt that this hand should be opened 4♠, but looking at the cards, I think it's too strong to be a preempt. I found this really surprising. Most people would not preempt with this hand because they don't know what level they should be in, but the distribution of the proper level is much much narrower than Hand A. Interesting... The reason turns out to be that many of our par slams are sacrifices, which is much more likely with hand A. Hand B is much more likely to make game without a sacrifice.

Hand C was guessed to be a two-suiter or something that suggested that maybe spades wasn't always the best suit. Here it is:
KQxxxx (Kxx xx xx)

A very down the middle "good weak 2." Again, I was surprised how often spades is not our best suit. I can imagine that if we had a different long suit besides the master suit, it would be even more often that we would prefer not to be in our long suit.

What was the point of this exercise? I'm not sure. But is was interesting to look at. Plus I wanted to show you how hard it was to use par data to come up with something meaningful about preempts.

[Blofeld]

tysen2k, on Oct 10 2005, 11:18 AM, said:

Blofeld, on Oct 7 2005, 04:11 PM, said:

My understanding is that "our par contract" means:

(1) the par contract, if it is ours
(2) the highest contract played by our side such that we would prefer it to the par contract, if the par contract is them making something
(3) our best contract, if the par contract is them sacrificing.

But that's a rather contorted definition. So perhaps my understanding is out.

That's it except for one small nit on point #2. Our par is the lowest possible sacrifice that's still profitable. Let's say we can take 11 tricks in hearts and they can take 11 tricks in spades. If you took your definition #2 literally, then it would say that our par is 7♥, when it should be 6♥.

I think my definition works, because [in your example] the par contract is 6♥x-1 by us, so point #1 applies and we never move onto #2.

---

Your data are interesting, but could you tell us where the thirteenth card in hand C should be?

I like the point that par contract is not best for determining preempts (unless this is going to shift people over to opening 4♠ on hand .

[tysen2k]

Blofeld, on Oct 10 2005, 10:06 AM, said:

tysen2k, on Oct 10 2005, 11:18 AM, said:

Blofeld, on Oct 7 2005, 04:11 PM, said:

My understanding is that "our par contract" means:

(1) the par contract, if it is ours
(2) the highest contract played by our side such that we would prefer it to the par contract, if the par contract is them making something
(3) our best contract, if the par contract is them sacrificing.

But that's a rather contorted definition. So perhaps my understanding is out.

That's it except for one small nit on point #2. Our par is the lowest possible sacrifice that's still profitable. Let's say we can take 11 tricks in hearts and they can take 11 tricks in spades. If you took your definition #2 literally, then it would say that our par is 7♥, when it should be 6♥.

I think my definition works, because [in your example] the par contract is 6♥x-1 by us, so point #1 applies and we never move onto #2.

You are right

Quote

Your data are interesting, but could you tell us where the thirteenth card in hand C should be?

Sorry, it's a 6322 hand. Corrected in original post.