[The following questions come from the same person, sent to us in one mail. We have split it up into five smaller parts, to make it easier to see which answer refers to which question, and we have called the parts Question 8a, Question 8b, etc.]
Let me say up front that I am a user of The Law but have had healthy skepticism of it and I am not an apologist for it.
My criticsm of the book is not so much with the analysis – but on presentation. As another person wrote and you posted on your site, you'll forgive me if I say that the first 125 pages of your book is a waste – of my (and I suspect 95% of any reader's) time – and damn frustrating. That could have been summarized in three short lines:
Also, you seemed to go out of your way to attack Larry Cohen (outwardly in the first part of the book and coyly regarding 4-card overcalls near the back). Doesn't make you guys look good at all. Not classy. Attack the idea, not the person, especially in public.
Regarding the presentation, we refer to our answer to question 4.
We are sorry if you think we have attacked Larry Cohen. I can assure you that we definitely didn't mean to. We have tried to discuss principles and theories, not persons, but since Larry has made a lot of claims in print, which can be shown to be false, we simply had to say so. Is that an attack?
I see no clear, simple, definition of Working Points. Where does it say: Working Points are such and such? This is how you figure them. Instead, I get (forgive me) rambling notions of Building Up AQxxxx vs Kxx to 10 working points for no other reason than you say so (yea, the J). The PRESENTATION is poor.
HOW WOULD YOUR ANALYSIS BE AFFECTED IF YOU ADJUSTED DOWNWARD FOR WASTED POINTS TO ARRIVE AT WORKING POINTS INSTEAD OF ADJUSTING UPWARD (AS YOU DO).
Wouldn't that be more in tune with our intuitive feeling that we want to eliminate wasted points from our calculation such as when we have Qx in an opponent's suit that has been bid and raised? But if you keep your upward adjustment approach, at the very least, may humbly suggest that you rename Working Points, INVISIBLE POINTS, cause that's what you're really talking about, both in terms of high cards that aren't there and length tricks, when absolute winning high cards are not there as in a AQJxx vs. Kx in a side suit. It would give the reader a much better feel for things.
Wait. Maybe I have a better idea:
Since nowhere in the book do I find a definition of Working Points, let me try one: WORKING POINTS = actual high card points + Invisible Points - Wasted Points.
Invisible Points = High card points in your long suits that you don't actually have but your suit will play like you have them or small cards in (semi)solid long side suits. Wasted Points = High card points you do not expect to help you win tricks.
We have written more on working points here on our site. Sorry if we haven't made ourselves clear enough on the subject in the book. Working points is the sum of the honors, or spot cards substituting for honors, which take tricks on offense. If we have, say ace doubleton opposite five small in a side-suit and have the time to ruff out the suit (it is 3-3) and get two useful discards, those two extra winners are worth 3 WP each.
Your definition of Working Points sounds good to us. If you read our book carefully, you'll find that this is exactly what we say – so how could we disagree!
I am confused.
Suppose we use the best judgement we can using WP and SST + experience + knowing that xxx in RHO's suit is bad + devine guidance, etc. What is the highest 'batting average' we can expect to attain in competitive situations? Suppose we got 3 out of every 4 or 4 out of every 5 correct? That wouldn't be too bad, would it? But that is what YOUR DATA suggest would occur if we simply followed The Law!
At least in about half the cases, and more. Let me make my case.
You say that 16 and 17 trump hands occur a combined total of 49.78% of the time (page 22). You make a big deal that on 16 trump hands, the Law is only right 44.1% of the time (p.32). But that is not the point. Your table on p. 32 also tells us that with 16 trump (a mandatory 8 and 8 situation you say) that the total number of tricks is 16 or greater 77.8% of the time. Doesn't that strongly imply that it would be winning bridge to bid '3 over 2' even with only 8 trump and virtually never let them play at the 2-level. Even if they can make 9 and we make 7, they have to double before it becomes painful (assuming NV – you don't have to be suicidal when vul).
When there are 17 trump, your table on p. 33 tell us it would be OK to bid "3 over 3" 72.3% of the time, since someone could make at least 9 tricks that often. Even with 18 trump which occur 15.65% of the time, your table on p. 34 says that following the law by bidding would be right 68.8% of the time. Still not bad.
If we take all three where there are 16, 17, and 18 trump (which occur a combined total 65.4% of the time), bidding-on based only on trump length would get you a good result an average of 73% of the time! My bidding decisions should only be this good! So, unless I have missed something terribly, it seems to me that the Law may not so terrible a guideline to follow after all.
Can SST+WP do better than 73%. If it can, you haven't proven it. You've used anecdotal examples. Why not go back and do the same simulations that you did for The Law and see if your way works more than 3/4 of the time. Shouldn't you have done this from the beginning? Why do we have 'Actual Data' to denounce the Law but only anecdotal examples to support SST and WP?
With the Law seemingly correct 73% of the time when it tells you to bid, you might say, OK, when the law tells us to bid, maybe we should listen to it, but what about when it tells us not to bid. In those cases, we'll be wrong way too often. You could say that, but in fact on page 102, you conclude exactly the opposite:
"In my observation, if The Law tells you to bid, you can find a valid reason to pass fairly often. If the Law tells you to pass, it is usually right to do so." Isn't that in direct contradiction of the data about the Law you present? So, as I say, I'm confused. I know you guys have been looking at this stuff for a long time and I haven't. I respect that. But: What am I missing?
We're glad you brought this up, because it gives us the chance to kill another of those Law myths. You say you get "a good result 73% of the time by following the Law". But is that so?
Just because there are, say, 18 trumps and 18 tricks, both sides having nine trumps, it doesn't automatically follow that you will achieve "a good result" by following the Law and contracting for nine tricks. Some of the time your side can take 8 tricks in hearts and they ten in spades. If they bid their cold 4, what good did your competing to 3 do? How can such a deal possibly be labelled "a good result"? When the truth is that it's indifferent, or even bad: if your bidding 3 helped them in bidding a game they wouldn't have had you been passive, it instead produced a terrible result.
And if your side is the one taking 10 tricks, what good is it in following the Law and contracting for nine tricks? We much prefer+420/620 (or +300/500 if they save) to +170. Don't you?
In his study in The Bridge World (we mention it in our Statistics section), Matt Ginsberg came to the same, false, conclusion.
Time to tell the truth:
Even when there are at least as many total trumps as total tricks, there is NO GUARANTEE that contracting for the same number of tricks as your side's trumps will produce a good result. Often it will give you an indifferent result or a bad one.
Suppose both sides are vulnerable and the opponents bid 4. Expecting 18 trumps, you follow the advice of "always bidding 4 over 4". You get doubled, run into two unexpected ruffs and concede 1100. Yes, the opponents could make 6 for 1430, but if it's pairs and nobody else has bid the slam, or if it's IMPs and your team mates only play game, does the fact that the opponents can make a small slam make your result any better? How can such a deal be called 'a good result' just because tricks and trumps were equal?
And if we take a more mundane example, where the opponents could have made their contract, for +110, is your going -100 such a hot result? At pairs it might be, but at IMPs the result is a wash. Yet another indifferent result – but the Law pats its back and says 'a good result for me!'.
And how about the case where you have ten spades and they eight cards in both minors? Even if there are 18 total tricks, what's 'good' is it in playing 4 down one, when you actually owned the hand in 3? Or go two down in 4 doubled when all the opponents could make was +130? In both cases, the deal was according to the Law, and still it was BAD to follow the Law.
Finally, there are the cases when you will make your contract but you would have been better off defending – either because there were fewer tricks than trumps (the remaining 27%) or that you will make an overtrick in your contract. Suppose the opponents bid 3. Your side has nine diamonds and competes to 3, making ten tricks. Then, if there were 17 total tricks, it would have been better for you to double them (+300 or +500) or even pass if they were vulnerable (+200). Once more the deal is included in the 'good cases', even though doing the opposite would have been better.
The most important thing in competitive auctions is to avoid going minus when you could have gone plus. Results like -100 instead of -110 aren't worth much. But +110 instead of -110 is, just like +100 instead of -100, or if the contract is doubled +200 instead of -500 or +300 instead of -300, etc, etc. Our method concentrates on what one side can do. Therefore, we are in a better situation to judge than the Law who looks at the total, then tries to guess how the tricks are divided.
Our rule of thumb for part-score battles is: "If your estimation says your contract will make, go ahead and bid it – to avoid defending when both contracts make. But if your estimation says your side will not make your contract, don't bid it – to avoid declaring when neither side makes their contract."
Your analysis of SST looks like an extension of Loser Count (with a dose of judgement thrown in) – but you don't say so or give credit to Losing Trick Count Analysis. Just as you say on page 130 and elsewhere, balanced hands with nothing constructive to ruff require extra high cards. So does Loser Count.
Example: As you know, a 4333 hand with 0 HCP has 12 losers; opposite another 4333 zero-count that would mean 24 losers. Replace one of the 4333 hands with a 5521 hand with 0 HCP – it has 9 losers for a total of 21 losers – a difference of 3. Opposite another 4333 hand, 4333 has an SST of 6 and a 5521 has an SST of 3 – a difference of 3. See what I mean? So maybe without meaning to or realizing it, your SST analysis is a rehash of Loser Count. That doesn't make your analysis bad, it just makes it not new.
The Losing Trick Count (LTC) is one of the best ways of valuing your cards, but it suffers from one serious flaw. It looks at one hand at a time, instead of both together. That is the BIG difference between our method and LTC. Suppose you have 5-3-3-2 distribution with all the spade honors. That is 8 losers. Your partner has 3-4-3-3 distribution with all heart honors. That is 9 losers. 24-(8+9)=7, but these two hands take 8 tricks. If we change the second hand to 3-5-3-2 we cut one loser and get to a correct result (8 tricks), but if we change it to 3-5-2-3, the LTC formula says 8 tricks, but now it is 9.
In these three cases, LTC predicts right only once, while our formula gets the correct result each time, since SST will be 5, 5 and 4, respectively – giving a prediction of 8, 8 and 9 tricks.
Even if our method discusses losers, just like LTC, it is not a copy. The concept is original, and it is new. Some of our (wellknown) readers have even called it revolutionary...
On page 101 you warn that if you have wasted points in the opponent's suit, (the example you give is Qx) you state very emphatically that this STRONGLY (your emphasis) indicates that partner will be minimum for his bid. You state this but don't explain why you believe this is so. Could you help? Is it so obvious that it needed no explanation and Mel just doesn't see the obvious?
On page 146 you show this hand,
A J x x x
K Q x x x
where you open 1 and they overcall 2 and raise to 3. You say there's nothing to guide you but judgement as to bidding 3 or not. For many years I have been telling my students about Mel's Compete Count for exactly these situations. It is a rudementary rule but it gets them in the right ballpark. It utilizes Loser Count. MCC says you figure out Losers by Loser Count and Subtract 1 – that tells you how many tricks you can contract for while going it alone. In this example, we have 5 losers by LC-1 so we have 4 estimated losers and therefore 9 estimated winners, so we bid 3. And as you say on page 147 it is based on the sensible wish that partner fits one of your suits. It also wishes that partner has a little something – 1 Cover Card, to use LC parlance.
Forgive us for not being more explicit. If our opponents compete for the contract, we can expect them to have some values. If we have wasted honors in their suit, it follows that their share of the points consists of other honors. Had we had two small in their suit instead of Qx, for instance, they could have had that queen. Now they have some other honor(s) to make up for their missing queen, which tells us that the chance that our partner has an unexpected useful honor is smaller than when we have nothing in their suit. It's not a paradox, even if it sounds like one. Better to have 11 HCP with two small cards in their suit, than 13 HCP with queen doubleton in their suit.
Mel's Compete Count is surely a good rule, both to teach and practice. Stick to it.
More questions and answers:
[ 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 ]