[dustinst22]

 TylerE, on 2013-May-22, 14:31, said:

So, I made this thread...

http://www.bridgebas...856#entry726856

One the example hand, out of 10,000 trials, it made game 41.62% of the time, which is well above the equity point, so that strongly supports just blasting game.

Is Double Dummy analysis really the best way to determine if we belong in game?

[JLOGIC]

 TylerE, on 2013-May-22, 14:31, said:

So, I made this thread...

http://www.bridgebas...856#entry726856

One the example hand, out of 10,000 trials, it made game 41.62% of the time, which is well above the equity point, so that strongly supports just blasting game.

It strongly suggests inviting game (which is what I think is the correct action). There is no way that game will be a good proposition on the subset of hands that partner is rejecting the invite with.

[ArtK78]

 TylerE, on 2013-May-22, 14:31, said:

So, I made this thread...

http://www.bridgebas...856#entry726856

One the example hand, out of 10,000 trials, it made game 41.62% of the time, which is well above the equity point, so that strongly supports just blasting game.

But you don't have to rely purely on the odds. You can choose a middle ground, such as transfer and invite or transfer and pass unless partner preaccepts.

This would result in getting to game on many of the hands which make game and avoiding getting to game on most of the hands that don't make game. I am sure that this would improve your odds AND YOUR IMP EXPECTATION.

Whether it is better to pass unless partner preaccepts or to invite over a non-preaccept is an issue. The invite over the non-preaccept gets you to more making games, but also results in more minus scores (either in game failing or in 3 failing). I would choose to not invite but to accept over a preacceptance, but it is probably a close call.

[rogerclee]

 TylerE, on 2013-May-22, 14:31, said:

One the example hand, out of 10,000 trials, it made game 41.62% of the time, which is well above the equity point, so that strongly supports just blasting game.

This is not an argument for blasting game when you have the option of inviting, this is a pretty basic fallacy.

What you should really be simulating is how often game makes opposite a hand that would not accept an invite, not any hand period.

[TylerE]

The thing with inviting, is that when you invite and the invite isn't accepted, you're about 40 pts a board worse off than if you just always passed 2. I've done some additional equity calculations.

Basically, at vuln, always passing 2 has an EV of 140.8, always bidding 4 an EV of 178.0, and being in 3 an EV of 102.7. If your invites are 100% perfect, your EV jumps up to 290, but being wrong at all, ever, hurts it quite a bit.

[PhilKing]

 TylerE, on 2013-May-22, 14:31, said:

So, I made this thread...

http://www.bridgebas...856#entry726856

One the example hand, out of 10,000 trials, it made game 41.62% of the time, which is well above the equity point, so that strongly supports just blasting game.

Does it?

If we were not allowed to invite, that would be true.

[dustinst22]

 TylerE, on 2013-May-22, 14:55, said:

The thing with inviting, is that when you invite and the invite isn't accepted, you're about 40 pts a board worse off than if you just always passed 2. I've done some additional equity calculations.

Basically, at vuln, always passing 2 has an EV of 140.8, always bidding 4 an EV of 178.0, and being in 3 an EV of 102.7. If your invites are 100% perfect, your EV jumps up to 290, but being wrong at all, ever, hurts it quite a bit.

How are you calculating the EV for being in 3?

i.e. Being in 3 across any 15-17 hand? If this is the case, obv the EV won't be great because you'll be missing many games. I think to calculate the EV for invite you have to set the parameters to being in 3 across a minimum hand that opener won't accept with. Also, as Justin pointed out, opener will accept on many different types of hands (17 always, 16 sometimes, prime 15's occasionally). So plugging in a point value is not going to be very useful.

[PhilKing]

 dustinst22, on 2013-May-22, 15:05, said:

How are you calculating the EV for being in 3?

i.e. Being in 3 across any 15-17 hand? If this is the case, obv the EV won't be great because you'll be missing many games. I think to calculate the EV for invite you have to set the parameters to being in 3 across a minimum hand that opener won't accept with.

I did a short non-dd sim (32 hands) which suggested that partner was able to accept/reject very accurately.

In my method, I can invite and still stop in 2♥, which is obviously ideal.

[Wwchang]

 dustinst22, on 2013-May-22, 14:41, said:

Is Double Dummy analysis really the best way to determine if we belong in game?

Separate from this question, is the question of how well double dummy results compare to single dummy results. At least in the case of 4!h contracts specifically, Pavlicek's stats suggest they conform very closely.

[Wwchang]

I have some calculations on the other thread that suggest that blasting is about as good as random acceptance of invites, but if acceptance of invites is better than random, then inviting leads to a better overall result (caveat: does not take into account opposition bidding, and the simulation that was run does not exclude superaccept-type opening hands).

[TylerE]

 dustinst22, on 2013-May-22, 14:41, said:

Is Double Dummy analysis really the best way to determine if we belong in game?

In this situation, I think it's pretty dead on.

In partscores, DD tends to be a bit friendly to declarer (Get suits right 100% of the time), but in slams tends to favor the defense a bit (If there is a lead to beat it, it _will_ be found, even if odd and non-obvious at the table)

[aguahombre]

 dustinst22, on 2013-May-22, 14:41, said:

Is Double Dummy analysis really the best way to determine if we belong in game?

It is, indeed...if partner doesn't know what an acceptance looks like.

[mycroft]

...and if our opponents are perfect defenders...

Okay, and if we guess the Q 100%, and, and. Anybody have a feel as to whether declaring below DD is more or less likely than defending below DD?

[aguahombre]

 mycroft, on 2013-May-23, 12:12, said:

...and if our opponents are perfect defenders...

Okay, and if we guess the Q 100%, and, and. Anybody have a feel as to whether declaring below DD is more or less likely than defending below DD?

At some point in our off-and-on Bridge history, we came to a conclusion about all this:

The theoretical expectations from perfect defense and perfect dummy play are all a big wash. If we make better (informed) decisions than the field we win, if we don't we lose.

Where we gain is in the knowledge that the bidding decisions were partnership decisions, some of our defensive errors were partnership errors, and we have avoided masterminding ---which annoys both of us.

[PhilKing]

 mycroft, on 2013-May-23, 12:12, said:

...and if our opponents are perfect defenders...

Okay, and if we guess the Q 100%, and, and. Anybody have a feel as to whether declaring below DD is more or less likely than defending below DD?

There is little need for guesswork, since Pavlicek has compared top-level play with DD play over a large sample.

The defence is closer to par after trick 1, but often drop a trick on the lead (this maybe slightly bogus, since the lead sometimes only costs if declarer is also DD).

[Wwchang]

 PhilKing, on 2013-May-23, 13:57, said:

There is little need for guesswork, since Pavlicek has compared top-level play with DD play over a large sample.

The defence is closer to par after trick 1, but often drop a trick on the lead (this maybe slightly bogus, since the lead sometimes only costs if declarer is also DD).

on 4♥ contracts specifically, DD tricks averaged 10.01 and actual tricks averaged 9.99 over 8699 contracts. DD made 68.65% of the time, actual made 68.35%.
The difference is much larger for 3♥ contracts. There, average tricks were 8.41 DD and 8.49 Actual, with 3♥ making 52.82, and only 48.46 DD.


http://www.rpbridge.net/9x29.htm

[wank]

 TylerE, on 2013-May-22, 14:31, said:

So, I made this thread...

http://www.bridgebas...856#entry726856

One the example hand, out of 10,000 trials, it made game 41.62% of the time, which is well above the equity point, so that strongly supports just blasting game.

you quote a statistic and draw a random conclusion from it.

[Free]

I would transfer and pass. I don't have any intermediates in ♥, I don't have a shortness, I don't have an Ace,...

I wouldn't call inviting a mistake, but a matter of style. However, when you would invite with this hand, South shouldn't accept it! When you invite this hand, it means you play light invites, so you should accept heavy. South isn't heavy at all.

[Free]

 TylerE, on 2013-May-22, 14:31, said:

So, I made this thread...

http://www.bridgebas...856#entry726856

One the example hand, out of 10,000 trials, it made game 41.62% of the time, which is well above the equity point, so that strongly supports just blasting game.

Some people should really learn how to interpret statistics, or they should ignore them all together...

[RunemPard]

Mr. George...may we have the full hand. I have blasted texas on these hands in the past, and the only good results have been when the opponents belong in a making contract or unlikely game. So I am going to go with transfer/pass.