[beatrix45]

I really hate to challenge something that bears the name of two legends of the game. That said, I keep seeing people quoting the K and R evaluation of a given bridge hand, and sometimes the result just does not make sense to me. The idea is often whether to upgrade or downgrade the ordinary Milton Work/Goren point count hand evaluation for purposes of making an opening bid. What's up with KnR?

[dake50]

Mostly it's pretending some particular feature is worth a point or two.
As if the accuracy of evaluation could be significantly bettered.
"Our evaluator is better than Goren" is their claim.
Succeeding better on a small group of hand cases that an expert would already know.
Neither statistically validated for frequency of "better",
nor decision tree (game or not) improved by how much?
Blowing smoke to sell books.

*** I know this attacks K+R disciples. Do the stats. Do the decisions.
Publish your results *IF* they support your claims.

[Cyberyeti]

I use it to challenge my own evaluation of a hand where partner's said something like "that wasn't close to an opening bid" as he does frequently, or somebody makes an evaluation on these forums that I think is a mile off.

I get a few surprises, but usually my assessment of "my 5431 10 count was a lot better than your 5422 11" agrees. It's an OK tool if not treated as gospel, there have only been 2 or 3 hands where I think it's way out.

[hrothgar]

dake50, on 2014-August-05, 06:03, said:

Mostly it's pretending some particular feature is worth a point or two.
As if the accuracy of evaluation could be significantly bettered.
"Our evaluator is better than Goren" is their claim.
Succeeding better on a small group of hand cases that an expert would already know.
Neither statistically validated for frequency of "better",
nor decision tree (game or not) improved by how much?
Blowing smoke to sell books.

*** I know this attacks K+R disciples. Do the stats. Do the decisions.
Publish your results *IF* they support your claims.

There have been any number of studies that demonstrate that Goren 4-3-2-1 point count significantly under values Aces and overvalues queens.

Low and behold, the core of K+R is a A=3, K=2, Q=1 valuation.
(There are a few other bells and whistles, but this is the most significant deviation)

Simply put, I don't think that the burden of proof is on my side.

BTW, your comment about blowing smoke to sell books is incredibly ignorant and completely misunderstands the history of CCCC

[jogs]

Cyberyeti, on 2014-August-05, 06:14, said:

I get a few surprises, but usually my assessment of "my 5431 10 count was a lot better than your 5422 11" agrees. It's an OK tool if not treated as gospel, there have only been 2 or 3 hands where I think it's way out.

That's what I think. 3-1 in the side suits is worth between 1 and 1.5 points better 2-2 in the side suits. It's worth a little more than 1/3 of a trick.

[dake50]

Simply put, I don't think that the burden of proof is on my side. -- hrothgar

*** I didn't at all think y'all would take the challenge to prove your theory. You didn't!!
*** Hooray!! You claim 4321 count is bettered, that's some claim even though you don't demonstrate it.

*** Are you possibly claiming Kaplan and Rubens DID NOT PROFIT FROM THEIR BOOKS SELLING??? Incredibly ignorant right back.
Just like Goren's 4321 made millions? But not them?

*** Please put me to the demonstrated stats/else that they validated their claims.

*** Like I said this'll rile up the devotees. You are riled up.!

[Siegmund]

Quote

Are you possibly claiming Kaplan and Rubens DID NOT PROFIT FROM THEIR BOOKS SELLING???

They did profit from people reading the Bridge World, I imagine.

But they did not push their count on people in all their books the way Goren did. Rubens wrote several excellent books, including some golden chapters on hand evaluation, without even mentioning "his" system.

If you want facts as to how different mechanical evaluation methods compare against each other, the material is out there. See, for instance, http://bridge.thomas...s/original.html which confirms Kaplan-Rubens is an excellent evaluator of offensive strength in a suit contract (but only fair for defensive strength or notrump bidding -- no method is a one-size-fits-all perfect evaluation system.)

[gordontd]

dake50, on 2014-August-05, 20:44, said:

*** Like I said this'll rile up the devotees. You are riled up.!

I see no evidence of this.

[MickyB]

K+R is either bad or just grossly misused. It has some value for suit contracts but none for no-trumps. I would never take it as an indication of what I should open - a 5-5 9-count can pass and catch up later, but it can't open then persuade partner to not bid 3NT with a misfitting 13-count.

[paulg]

MickyB, on 2014-August-06, 02:14, said:

K+R is either bad or just grossly misused. It has some value for suit contracts but none for no-trumps. I would never take it as an indication of what I should open - a 5-5 9-count can pass and catch up later, but it can't open then persuade partner to not bid 3NT with a misfitting 13-count.

A young man is aging before my eyes.

[rhm]

As others I have my own evaluation method based on double dummy studies.
K&R was developed before these studies became available.
And as others have already pointed out K&R is quite accurate for suit contracts but not for notrump contracts with balanced hands, where simple standard works at least as good.

In other words I think it is much better than standard for unbalanced hands

Rainer Herrmann

[rhm]

MickyB, on 2014-August-06, 02:14, said:

I would never take it as an indication of what I should open - a 5-5 9-count can pass and catch up later, but it can't open then persuade partner to not bid 3NT with a misfitting 13-count.

There is some truth in this and lot of fallacy.
The theory is with distributional assets you can pass and can catch up later.
Meanwhile you will not end up in a hopeless 3NT when the hands do not fit.

I do not share this optimism
Bidding like play is more often than not a matter of probabilities and frequencies, not one of perfection.
If I believe that my chances having a fit is great, which usually means opponents will also have one, considerations whether 3NT will have play will not hinder me.
In bidding tempo and time matters. It is often a contest, which side gets more and better information first, the latter being particularly important with distributional assets.
Avoiding hopeless contracts is not the only concern. Reaching good ones without having to guess after opponents strike first possibly with a preempt is at least as important
Passing first and catching up later has hidden costs.

Rainer Herrmann

[Zelandakh]

As hrothgar points out there is plenty of evidence to suggest that Queen Points (3/2/1) and the Milton equivalent (4.5/3/1.5/(1)) provide more accurate evaluation for suit contracts than unadjusted Milton. It is also the case that that downward adjustment for queens is not constant - an unsupported queen needs a big adjustment whereas a queen supported by a higher honour in the same suit is close to full value. That leads to my suggested evaluation method for intermediates, in which the basic Milton count is adjusted upwards by +0.5 per non-singleton ace and dowanwards -0.5 for each unsupported quack. Honours in short suits should also be adjusted downwards (eg sgl K = 2 is a reasonable compromise for a notoriously difficult to evaluate holding).

The other big point about KnR was made by Micky, it is often misused. A classic example is a hand where we already showed a fair degree of distribution and then use a high KnR evaluation to justify bidding on despite the case that much of that evaluation is already known and would be factored in for any hand that we could still hold. Despite that it remains a useful check tool sometimes. You just need to be careful about what it is telling you and sometimes checking against some control hands helps in interpreting the scores.

[uday]

This site has a lot of really interesting reading on hand evaluation if you got some time to be vacuumed up http://bridge.thomas...com/valuations/

[AreyHakaal]

Taken IN ISOLATION if we take percentages/proportions of trick taking ability of honor cards then an Ace will win 100% ,a king will win 50 % when the ace is favorably placed!a Queen will win at the most 25 times when both AK are well placed.So the correct point count is 4210. the Jack in isolation will never win.However during the auction these values go on changing as we all know very well.The point count method which ever great masters have given is only a rough guide for the new learners of Bridge.the HCP and the losing tricks count(LTC) as also the top winners have all to be taken in account when we are right at the table.

[WellSpyder]

AreyHakaal, on 2014-August-07, 01:57, said:

Taken IN ISOLATION if we take percentages/proportions of trick taking ability of honor cards then an Ace will win 100% ,a king will win 50 % when the ace is favorably placed ...

There are 4 possible hands the A can be in, not 2. If it is in the same hand as the K or in the partner's hand then the K will normally be worth a trick, as well as when it is in RHO's hand. So the K will be worth a trick roughly 3 times out of 4, and if you value an A at 4 points then you might value a K at 3 points....

[Zelandakh]

AreyHakaal, on 2014-August-07, 01:57, said:

Taken IN ISOLATION if we take percentages/proportions of trick taking ability of honor cards then an Ace will win 100% ,a king will win 50 % when the ace is favorably placed!a Queen will win at the most 25 times when both AK are well placed.So the correct point count is 4210. the Jack in isolation will never win.

This is the logic behind Culbertson's original HT evaluation method - ace = 1, K = 0.5, Q = + ~= 0.25. However, as WellSpyder points out this is not optimal. Assuming we are not looking at the ace ourselves the king is going to fetch a trick about 2/3 of the time, if the ace is with partner or RHO. That gives the more familiar 3:2 or 4.5:3 ratio already given. At NT the king is going to be a little better than this though, since sometimes the defenders need to attack the suit from the wrong side in order to avoid losing a tempo. That is one reason why the 4:3 count works better at NT than for suit contracts.

[AreyHakaal]

WellSpyder, on 2014-August-07, 03:31, said:

There are 4 possible hands the A can be in, not 2. If it is in the same hand as the K or in the partner's hand then the K will normally be worth a trick, as well as when it is in RHO's hand. So the K will be worth a trick roughly 3 times out of 4, and if you value an A at 4 points then you might value a K at 3 points....

Please read my full analysis.When one is making the FIrst Bid ,be it any denomination,one does not know where the Ace of a suit will be.Neither is there any reason to suppose that 33.33 % it will be with partner. I have calculated the winning possibility when one presumes that the Ace is with the opponents and all the further logic is on the same basis.And I have even said that the trick taking value changes and so does the HCP value of a honor card when the bidding proceeds further and the experts always revalue the HCP value of any honor card as indicated by the opponents silence or otherwise.I am afraid ,you have based your valuation only on mathematical probability.I am giving a hand,you hold,xxxx,AQxxx,void,xxxx,and RHO opened a weak Notrump .My partner with the given hand called 2 C (Landy).I held x,xx,xxxx,AKQxxx and passed.Opener bid 2diamond and my P bid 3 Clubs, His LHO bid 4 Spade and I bid 5 clubs And the HCP counting opponents doubled so loudly that the roof of an adjoining house was blown away .Result was one over trick.This hand was reported in the bulletin of the state level tournament.The 25 HCP with opponents produced ONLY once trick.That's why one has to be right at the table.

[Zelandakh]

AreyHakaal, on 2014-August-07, 05:54, said:

Please read my full analysis.When one is making the FIrst Bid ,be it any denomination,one does not know where the Ace of a suit will be.Neither is there any reason to suppose that 33.33 % it will be with partner. I have calculated the winning possibility when one presumes that the Ace is with the opponents and all the further logic is on the same basis.

Given we are talking about eveluation in a vacuum away from any specific hand for the the time being, there seems to me to be a much better reason to suppose the ace is with partner a third of the time than 50% left, 50% right and 0% opposite. One of the things I learned in analysis class is that given the right set of base conditions one can prove pretty much anything with further logic. That is why getting the right base conditions (axioms) is regarded as of fundamental importance.

[Lovera]

AreyHakaal, on 2014-August-07, 05:54, said:

Please read my full analysis.When one is making the FIrst Bid ,be it any denomination,one does not know where the Ace of a suit will be.Neither is there any reason to suppose that 33.33 % it will be with partner. I have calculated the winning possibility when one presumes that the Ace is with the opponents and all the further logic is on the same basis.And I have even said that the trick taking value changes and so does the HCP value of a honor card when the bidding proceeds further and the experts always revalue the HCP value of any honor card as indicated by the opponents silence or otherwise.I am afraid ,you have based your valuation only on mathematical probability.I am giving a hand,you hold,xxxx,AQxxx,void,xxxx,and RHO opened a weak Notrump .My partner with the given hand called 2 C (Landy).I held x,xx,xxxx,AKQxxx and passed.Opener bid 2diamond and my P bid 3 Clubs, His LHO bid 4 Spade and I bid 5 clubs And the HCP counting opponents doubled so loudly that the roof of an adjoining house was blown away .Result was one over trick.This hand was reported in the bulletin of the state level tournament.The 25 HCP with opponents produced ONLY once trick.That's why one has to be right at the table.

This hand will been. played probably with crossed ruff because this points don't take so many tricks and in my valutation this plain gets off controll.