Long Rambling Rant Warning. Looking at upgrade/downgrade ideas....
AQ75, K64, KQJ, K52
It is interesting that some (at least justin and han) consider this hand a"good 18" because it has an ace and three kings. Good enough to overcome the poor distribution (4333). Perhaps they are right. Certainly in Justin's case and perhaps in han's they take more tricks with 18 hcp than I do,so they don't need to be as cautious when downgrading hands. But I want to consider what factors influencea decision to upgrade (better than average) or downgrade (worse than average) a hand.
You might think we should not be talking about upgrading this 18 hcp (after all it has more than the maximum for 1NT opener) so only downgrade values that should be considered. However,
I automatically downgrade all 4333 handpatterns, and not just for opening 1NT. So for me NOT to open 1NT on a 4333 hand with 18 hcp, it would have to have some feature to make me
upgrade it for the downgrade came automatically with this hand pattern.
What features might cause someone to upgrade an 4333 hand? Well, clearly controls are one feature. Both Justin and Han mentioned this hand has an ace and three kings (5 controls). Another is the presences of Ten's and nine's (the more the merrier). Location of honors in combinations is another as well as honors in the long suit (for example AQxx Kxx KQJ Kxx is better than KxxxAQx KQJ Kxx because of two honors in the long suit in the first case).
But how rare is FIVE controls when holding 18 hcp and a balanced distribution.The maximum is eight controls (something like AKx Kxxx Axx Ax) and the most frequent holding seems to be six controls can be gotten to several ways (AxxAJxx Axx AJx, AQxx Axx Axx Axx, AKx Kxx. Kxx. KQxx, AKx Kxx KJx KJx). The fewest controls you can have with 18 is two (specifically two kings and all four queens and jacks, so that is very rare).
So one question might be, is 5 controls a lot for a flat 18 hcp hand? That is, is it relatively uncommon a holding with the average considerably less than 5 controls?
One way to approach this is to manually examine a lot of hands. The hands I printed out from some old JEC matches on the first page of this thread showed this frequency for controls
From JEC match posted earlier
Controls Frequency
2 0
3 0
4 1
5 10
6 22
7 14
8 3
This suggest that 5 controls is a "bad" 18 hcp in the context of what is usually held by hands with 18 hcp. Of course a better way is to use math to determine the expected frequency of hands with so many controls for 18 hcp, or to use a large database of hands dealt and played to estimate the expected frequency. I combined dealer hands with 18hcp and 4333 distribution across five Bridgebrowser databases to get a collection of 106,954 played hands which worked out to be 6366 unique deals. When you break down thedeals the ratio of hands per control was (normalized to 5 controls)
Cntrls Normalized
2 0.000806
3 0.013999
4 0.179365
5 1
6 2.087943
7 1.563372
8 0.538245
So it turns out, 5 controls is less than the expected number of controls for a hand with 18 hcp. The average number of controls is actually 6.22 controls (based upon this bridgebrowser data) So as far as
upgrading, a good 18 might be more than 5 controls with 5 controls below average for 18 hcp.
One speculative evaluation method is Banzai points as recommended by Ron Klinger. The original hand in this post evaluates out as 25 Banzai points, and Klinger suggest opening 1NT with 22 to 25 Banzai points.Adding just one TEN to the hand would make it too strong for a Klinger 1NTopening bid. Interesting, swapping the ACE for two queens would reduce the controls to 3 but but would increase the Banzai points to 27. Would that simple change make the hand stronger or weaker? In Klinger's view, it makes the hand stronger.
In the real world, that is an interesting question.
The hand AQ75, K64, KQJ, K52 (25 Banzai points, five controls) would changeto something like
QJ53, KQ6, KQJ, KJ2 (27 Banzai points, 3 controls).
Interesting, if you put the Control rich (8 control hand) into the mix, it has even less Banzai points. (AKxx Axx Kxx Axx) = 18 hcp butonly 23 Banzai points, close to a minimum 1NT opening bid using Banzai as your "method."
So this spurred me to ask the following question. Does the average result of opening a 4333 balanced hand change with number controls (with respect to a 1m or 1NT opening bid)? To test this I combining a lot ofbridge hands from multiple online play, creating a new database of 106,954 hands where the dealer had 18 hcp and specifically 4333 hand pattern. Next, I probed it based upon the number of controls held. Lets start with the overall results for opening 1m or 1NT (for the sake of argument, we will forget about the other possible openings, so things like Mexican 2
♦ and four card majors are not included in any of the following, but those bids do have some influence on the averages of the other bids.
First, for all 18 point hands lumped together.
Bid Deals AvIMP # Imp AvMP #MP
1C 33066 -0.03 31126 49.84 1940
1D 13822 -0.11 13110 47.91 712
1H 777 -0.81 758 47.94 19
1S 727 -0.93 706 54.13 21
1NT 56584 0.12 54992 51.27 1592
2C 143 -1.51 141 63.34 2
2D 69 -0.67 63 53.74 6
2H 11 -1.46 11 50 0
2S 4 -5.4 4 50 0
2NT 1245 -0.94 1208 49.95 37
3H 1 5.6 1 50 0
3S 1 -5.8 1 50 0
3NT 91 -2.19 90 37.5 1
4C 1 -12.27 1 50 0
4H 1 -5.13 1 50 0
4S 1 -9.73 1 50 0
4NT 3 -5.38 3 50 0
5NT 3 -7.27 3 50 0
Next lets compare results for imps for 2 to 8 controlsand only opening 1m or 1NT. To simplify the comparison, I averaged the resultsfor 1♣ and 1♦ opening bids for the comparison (taking into account thenumber of bids and the average score for the number of bids by this equation.
[/font][/size]
1[cl] bids x average 1[cl] result, + (1[di] bids x average 1[di] results)
1m average = ____________________________________________________________________
(total 1[cl]+1[di] opening bids)[/font][size="3"]
Cntr 1mimp 1N imp 1m hands 1N hands
2 -2.42 0.80 4 12
3 -0.60 0.48 111 163
4 -0.17 0.20 1365 1878
5 -0.095 0.15 8170 10289
6 -0.084 0.15 17237 21650
7 0.005 0.08 12793 15697
8 0.057 0.01 4536 5285
Understanding the limitation of looking at data from which this is extracted is important. First some pairs still play 16-18 notrump, others play precision or similar forcing club. The normal 16=18 peoplehave no problem worrying about what is being thought/discussed here. Theprecison guys can open 1C and rebid 1NT easily enough. So drawing too much of aconclusion is risky. But one thing seems clear, as controls increase(especially to 7 or 8), opening 1NT becomes less advantageous and open 1m lessdisadvantages as reflected in the average scores. DO NOT read anything into the2 control and 3 control data, as there are far too few hands played in the current data to reflect useful information.
A couple of things I did note.
1) upgrade with SEVEN or EIGHT controls seems right. Also, it is 7 or 8 that is "more than" the average 18 hcp hand holds
2) Banzai points were not supported. That is, an eight control hand has the fewest Banzai points but the data suggest opening 1m-then rebidding 2NT is a better choice than downgrading and opening 1NT. Also, despite the small sample size, the 2 and 3 and 4 control hands tend to be richer in Banzai points and yet those were the worse for treating as too strong for 1NT.
3) anyone drawing too much out of this is crazy. However, I have the database of hands if someone wants to propose interesting questions, say, how does the different 18 point hands average opposite a partenr with 8 hcp? Etc.
Help
jillybean, on 2011-July-17, 14:29, said:
s
t
r-m
nd
ing) tr. v. - Any bid made by bridge player with which partner disagrees.