Invite or Bash?

[Fluffy]

I am not the least bit surprised that a simulation shows that bidding is the right action. I appreciate that the simulation plays double-dummy, but remember that it also defends double dummy, and will always make the killing lead if there is one. Defence is the toughest part of the game and on auctions like this one, the lead is often a very difficult choice, where choosing wrong dooms the defenders, while often providing a lot of inferential information to the declarer, not to mention tempo and even an outright trick on occasion.

Everything you said is right on most hands, however on this hand, were we have two or three 4-3 fits it favours a lot declarer, and in my opinion even more.

 

Anyway the point I wanted to make from the start of this hand is that on this kind of 3NT contract I'd rather hold QJxx than Kxxx, I wonder if others agree on this.

[kenberg]

I am one of the 2NT bidders. I ask: If we change it to

 

 

[hv=b=7&d=n&v=-&n=sa752hk842dktc752&a=pp1n(14-16)p2cp2dp?]166|300[/hv]

 

is this better or worse? I think better.

 

The role of spades is to guard the suit, I hope for tricks from the minors. With the above change, I go to 3NT.

 

I changed the spade T to a spot and the diamond spot to a T.

[nige1]

[hv=b=7&d=n&v=-&n=sat75hk842dk5c752&a=pp1n(14-16)p2cp2dp?]166|300|

 

32-board team match, IMPs.

This would be easy if partner had replied 2M, or if you were vulnerable... [/hv]

IMO 3N = 10, 2N = 8. I appreciate Fluffy's arguments but you have...

• Combined 24+ HCP.
• Good quality working points mostly in your long suits.
• No weak doubleton.
• Useful shape (4423).
• Fair intermediates (T877)

[Mbodell]

If we change it to ...

 

I had a similar question of how close it was and what we change it to. So I decided to see what the effect of turning the spade T into the spade 9. Note, some ~25% of the time there should be very close to no effect because any hand in the simulation that used to put a spade 9 in partner's hand will now be putting the spade T instead, but ~75% of the time the spade T will be with the opponents (where the 9 used to be). Before giving the numbers, I reran a 100,000 sample of the original hand to help people get a feel for what differences to expect just in different samples (as opposed to different hands). The new sample (with the exact same hand- I.e., still have the ST in the numbers below) gave:

 

There were an average of 8.35278 tricks available in nt

In general, blasting instead of inviting wins you 0.07939.

In general, inviting instead of blasting wins you -0.07939.

=================================================

When we stay low we would have made game 42273 times (42.272999999999996%)

When we stay low we would make 8 exactly 43415 times (43.415%)

When we stay low we would make 7 or fewer exactly 14312 times (14.312%)

 

This is really very close to before (as expected). the average number of tricks is within 0.001. The IMP differential is 0.00031. The percentage numbers are within about 1/10 of one percent. All well and good.

 

In contrast, with the spade 9 instead of T, we get:

There were an average of 8.18232 tricks available in nt

In general, blasting instead of inviting wins you -0.6785.

In general, inviting instead of blasting wins you 0.6785.

=================================================

When we stay low we would have made game 34050 times (34.050000000000004%)

When we stay low we would make 8 exactly 46750 times (46.75%)

When we stay low we would make 7 or fewer exactly 19200 times (19.2%)

 

So we've lost about 0.17 tricks on average (somewhere around 1/6 of a trick). Game is about 8% less likely, making exactly 8 tricks is about 3% more likely, and 7 or fewer is about 5% more likely. This transformation makes this quite bad to blast. -0.7 IMP (recall that the T instead of 9 gave +0.07 IMP - note the extra 0!).

 

Again all the usual caveats about double dummy and is it good or not, but you can see changing the T to the 9 makes a huge difference on this hand, at least in double dummy world. The modified script is here.

[Fluffy]

Don't stop there just now! upgrade it to the jack :)

[whereagles]

If I recall correctly, double dummy tends to benefit defenders on part scores and declarer on slams. At game level it's about break even.

[BillPatch]

I am one of the 2NT bidders. I ask: If we change it to

 

 

[hv=b=7&d=n&v=-&n=sa752hk842dktc752&a=pp1n(14-16)p2cp2dp?]166|300[/hv]

 

is this better or worse? I think better.

 

The role of spades is to guard the suit, I hope for tricks from the minors. With the above change, I go to 3NT.

 

I changed the spade T to a spot and the diamond spot to a T.

The ♦10 is only useful if partner has abnormal length or a combination of high cards associated with abnormal length. Since the mode for the length of the opener's diamond suit is 4, it is very likely that the diamond T is wasted. Since the mode for opener's spade length is 3, and that T is in a longer suit, it is more likely that the ♠10 is better.

[kenberg]

I thank folks for commenting on my variation. I ws thinking that when opener does not have a four card major them fairly often he has a five card minor. I have no serious estimate how often this holds when the auction begins 1NT-2C-2D. But the fact that weakening the spde T to even the 9 causes a real problem gives me pause.

 

Anyway, thanks. And apologies to the OP for messing with his cards, I just got curious.

[whereagles]

by the way, what's up with so many significant figures in the calcs above? It's not very meaningful :)

[Fluffy]

I thank folks for commenting on my variation. I ws thinking that when opener does not have a four card major them fairly often he has a five card minor. I have no serious estimate how often this holds when the auction begins 1NT-2C-2D. But the fact that weakening the spde T to even the 9 causes a real problem gives me pause.

 

Anyway, thanks. And apologies to the OP for messing with his cards, I just got curious.

People sometimes accept invites with minimums with 5 card minor.

[Mbodell]

by the way, what's up with so many significant figures in the calcs above? It's not very meaningful :)

 

The output is just what the program outputs as raw results. The quoted calculations are for 100,000 hands, but the same code works for 10 hands or 1000 hands or 1,000,000,000 (if you have enough time), so I usually don't round off and let myself and others figure out how many digits are significant.

[Mbodell]

Ok, so I ran a few other scenarios of 100,000 hands (it takes about 10 hours for each of these simulations, although I can run multiple in parallel, so that's why it takes me a while to sim people's hand requests).

 

First the hand exactly how Kenberg had it. This is A752 K842 KT 752. The results are:

 

There were an average of 8.17807 tricks available in nt

In general, blasting instead of inviting wins you -0.68217.

In general, inviting instead of blasting wins you 0.68217.

=================================================

When we stay low we would have made game 34036 times (34.036%)

When we stay low we would make 8 exactly 46835 times (46.835%)

When we stay low we would make 7 or fewer exactly 19129 times (19.128999999999998%)

 

That is much more similar to the turning the spade T into the 9, and makes the invite better than the blast. At least according to double dummy running. So this matches people's idea that the T in the short diamond suit, even though partner might have length there, is not as valuable as in our 4 card spade suit.

 

I then wondered if part of the problem was Kenberg's treatment also weakened our spots. The spade T became the diamond T but the diamond 5 became the spade 2. So I decided what happens if you keep the spot strength the same, just move things around. Take the original hand and spade T becomes diamond T, diamond 5 becomes heart 5, and heart 4 becomes spade 4. This gives you a hand of A754 K852 KT 752 as at here. As compared to the original AT75 K842 K5 752 both have the same 1 T, 1 8, 2 7s, 3 5s, 1 4, and 2 2s.

 

There were an average of 8.17226 tricks available in nt

In general, blasting instead of inviting wins you -0.70255.

In general, inviting instead of blasting wins you 0.70255.

=================================================

When we stay low we would have made game 33785 times (33.785%)

When we stay low we would make 8 exactly 46845 times (46.845%)

When we stay low we would make 7 or fewer exactly 19370 times (19.37%)

 

It is unsurprising that this is virtually the same as the Kenberg hand as all we've done is promoted A752 => A754 and promoted K842 => K852, with the other two suits the same. It is marginally surprising that the hand with the weaker spots took more tricks and scored better, but that is just sample fluctuations.

 

Lastly, Fluffy asked/joked about changing the ST to the SJ in the original hand to see what that would do. Remember changing the ST to the S9 dropped about 1/6 of a trick on average, and lowered our gain from blasting by about three-quarters of an IMP (from +0.07 IMP to -0.68 IMP). How much better is a J than a T?

 

There were an average of 8.71876 tricks available in nt

In general, blasting instead of inviting wins you 1.7388.

In general, inviting instead of blasting wins you -1.7388.

=================================================

When we stay low we would have made game 59080 times (59.08%)

When we stay low we would make 8 exactly 32920 times (32.92%)

When we stay low we would make 7 or fewer exactly 8000 times (8.0%)

 

So the J over the T is worth more than 1/3 of a trick on average, as this is about +0.366 tricks (consider that 1 hcp = 0.325 tricks if 40 hcp = 13 tricks, so this is a little better than expected by strict hcp). By IMPs we go from blasting being only quite marginally positive to a huge +1.7 IMP positive. Unsurprising given we go from making game only around 42% of the time to 59% of the time! So a J above a T is about or slightly more than twice as good as a T over a 9. But, surely, everyone would know to blast game with this hand as an 11 count opposite 14-16.

[Fluffy]

Not joking at all Michael, A jack combined with an ace is a strong holding, and double dummy perhaps even more. I think you are wrong that it is woth 0.366, you are comparing it with the T, while for calculating what 1 HCP is worth IMO you should compare it with a blank or the 9, there it will be a lot more (About half a trick difference).

 

I expected the Jack over 10 ratio to be similar to the 0 over 10, 2 times the difference not really expected.

[Zelandakh]

Could I make another request of you please Mbodell. Could you run the same sim with hands of ♠A752 ♥K842 ♦K5 ♣J75 and ♠A752 ♥K842 ♦K5 ♣T75. That is, moving the minor honour to be unsupported by a higher one. Cheers mate. :)

[kuhchung]

Could you also get me a beer while you're at it Bodell?

 

:)

[Mbodell]

Alright, some more hands:

 

Everything you said is right on most hands, however on this hand, were we have two or three 4-3 fits it favours a lot declarer, and in my opinion even more.

 

Anyway the point I wanted to make from the start of this hand is that on this kind of 3NT contract I'd rather hold QJxx than Kxxx, I wonder if others agree on this.

 

I simmed AT75 QJ84 K5 752 instead of the original AT75 K842 K5 752. So the only change is K842 => QJ84. A number of people liked forcing because of the AKK nature, but Fluffy thought QJxx was better than Kxxx. Banzai points would agree with QJxx worth 5 to the 4 for Kxxx. Results:

 

 

There were an average of 8.35872 tricks available in nt

In general, blasting instead of inviting wins you 0.22545.

In general, inviting instead of blasting wins you -0.22545.

=================================================

When we stay low we would have made game 43684 times (43.684%)

When we stay low we would make 8 exactly 42309 times (42.309000000000005%)

When we stay low we would make 7 or fewer exactly 14007 times (14.007%)

 

Double dummy agrees that QJxy is slightly better than Kxy2 for a suit.

 

Could I make another request of you please Mbodell. Could you run the same sim with hands of ♠A752 ♥K842 ♦K5 ♣J75 and ♠A752 ♥K842 ♦K5 ♣T75. That is, moving the minor honour to be unsupported by a higher one. Cheers mate. :)

 

So two hands asked for here. First for the T in clubs, not spades, we simulate A752 K842 K5 T75. Results:

 

There were an average of 8.15626 tricks available in nt

In general, blasting instead of inviting wins you -0.90265.

In general, inviting instead of blasting wins you 0.90265.

=================================================

When we stay low we would have made game 32006 times (32.006%)

When we stay low we would make 8 exactly 48771 times (48.771%)

When we stay low we would make 7 or fewer exactly 19223 times (19.223000000000003%)

 

So the T in the clubs quite a bit worse than the T in spades. In fact the T in clubs is worse than the T in diamonds and worse than the 9 in spades!

 

What about the J of clubs? I also did A752 K842 K5 J75 with results:

 

There were an average of 8.51464 tricks available in nt

In general, blasting instead of inviting wins you 0.66784.

In general, inviting instead of blasting wins you -0.66784.

=================================================

When we stay low we would have made game 49020 times (49.02%)

When we stay low we would make 8 exactly 41792 times (41.792%)

When we stay low we would make 7 or fewer exactly 9188 times (9.188%)

 

We've added about 1.5 IMP to the expected score, close to the 1.7 IMP we were adding for the spade T to spade J. It is better to have the club J than the spade T, but nowhere near as good as the spade J. We are also adding the same sort of 0.36 tricks or so when going from T to J.

 

Could you also get me a beer while you're at it Bodell?

 

[hv=pc=n&s=s76432hk5dakq7c62&n=sakqhaq32d62cakq3&d=s&v=b&b=7&a=1sp2cp2np7nppp]266|200|W has 1 spade, East has 2 clubs, both players follow to 3 rounds of hearts, and no one unguards a black suit or throws the 13th heart.[/hv]