[the hog]

 jogs, on 2014-February-18, 11:54, said:

Actually due to the importance of trump length, I think 4-card majors is passé. 5-card majors makes it easier to bid slams. Identify those 5-4 fits quickly.

Sorry, this is hijacking this thread, but a comment like Jogs' needs a reply.
Actually I think 4 card Ms is better. It is more pre emptive and lets you find 4-4 fits much more quickly. Look at the Hacketts and Auken von Arnim, (though the latter play 4 CMs in a BC context).
The popularity of 5 card Ms is that most players are familiar with them and they are taught to beginners.

[Zelandakh]

 jogs, on 2014-February-18, 11:54, said:

Actually due to the importance of trump length, I think 4-card majors is passé. 5-card majors makes it easier to bid slams. Identify those 5-4 fits quickly.

Another unsupported and unsubstantiated assertion. Are you trolling? I take it from your silence that your previous post about Game Theory was hot air and simply trying to make your position sound "scientific" somehow.

The truth is that both 4 and 5 card majors have pros and cons and it depends on the rest of the system as to which works best. Identifying 5-4 fits quickly is great but we are not exactly going to be missing those often. Identifying 4-4 fits immediately and pressurising the opponents is also valuable. Then there is the half-way house of 5♠4♥ (Swiss Acol, etc) to throw into the mix. Your statement above is misleading because it ignores the complexities of the issue - unless you have something new to add to the debate (which has been going on longer than I have been alive).

[Cyberyeti]

 the hog, on 2014-February-18, 20:29, said:

Having played Acol for many years, I totally concur with this statement. Bidding 1NT is absurd with this hand. It certainly shows a stopper. We always showed the strong NT with a X in this position. I see no reason not to do it on this hand.

My choice at the table was between 1N and a slow pass as I had to think about it. I'm convinced now that playing what I play pass is the right bid (although I'm still not sure about the 17 count case) but have an aversion to slow passes in situations where partner may well have a marginal bid, so bid 1N.

[jddons]

I also held N but at my table W opened 2S weak (think opp forgot to mention an initial pass). So with the North hand, what's your poison? A double without hearts, 3C without a 6th club, pass with a much stronger hand than partner might expect or 2N without a stop! (wouldn't have thought of this without the previous discussion!)

[johnu]

 Cyberyeti, on 2014-February-19, 04:01, said:

My choice at the table was between 1N and a slow pass as I had to think about it.

There's also the choice between 1NT, slow 1NT, very slow 1NT

[helene_t]

 jddons, on 2014-February-20, 11:04, said:

I also held N but at my table W opened 2S weak (think opp forgot to mention an initial pass). So with the North hand, what's your poison?

3♣. WTP? I have a good suit.

[jogs]

 Zelandakh, on 2014-February-19, 03:26, said:

Another unsupported and unsubstantiated assertion. Are you trolling? I take it from your silence that your previous post about Game Theory was hot air and simply trying to make your position sound "scientific" somehow.

Yes, you are right. I'm actually using only ANOVA in most auctions. Nearly all bidding systems are point count based. They should be thinking more in terms of trick taking.

Expected(tricks) = trumps + (HCP-20)/3 + e(rror)

For trumps =4+4 and =5+4 the error is a normal curve with a mean of zero and a standard deviation of approximately 1.25.

The extra trump on average generates one additional trick.

Expected(tricks) = trumps + (HCP-20)/3 + SST + e(rror)

When you add Lawrence/Wirgren's short suit totals, you can improve those estimates. The std dev drops to 1 for flat patterns.

[jddons]

 helene_t, on 2014-February-21, 03:31, said:

3♣. WTP? I have a good suit.

thanks for your reply. the problem was centre hand opponent who decided the 5 level was where we wanted to be. then blamed me for bidding 3C!

[Cyberyeti]

 jddons, on 2014-February-23, 04:45, said:

thanks for your reply. the problem was centre hand opponent who decided the 5 level was where we wanted to be. then blamed me for bidding 3C!

TBF the 5 level is where you want to be but you are defeated by the 4-0 diamond break and 4-1 club break. You were however playing in the wrong suit, but from the right hand.

[Zelandakh]

 jogs, on 2014-February-21, 14:25, said:

Expected(tricks) = trumps + (HCP-20)/3 + e(rror)

You post this formula regularly and for me it is no different from your Game Theory reference earlier, an attempt to make your point of view sound "scientific". Note the "+e" component is redundant in the context of a hand evaluation and only adds "effect" (E for effect). To put it in context, what your eveluation scheme say is that you need 26hcp to make game with an 8 card fit and each additional trump is worth 3hcp regardless of the rest of the distribution including shortages. I know of no evidence to suggest this is more accurate than any of the alternatives, nor that it has any practical use at the table.

[jogs]

No, with a =4+4 fit you should be bidding game with 25 HCP. You need expected tricks at over 9 1/2 to bid game.

The std dev of the 'e' is critical. Under the old model which ignores trump length, that std dev is between 1 1/2 and 2.

E(tricks) = 6 1/2 + (HCP-20)/3 + e

By using two variables the std dev of the model drops to 1.25 to 1.5.


With a =5+4 fit you should be bidding game with 22 HCP.

Bring in the third variable, sum of the side suits total, you improve your estimates even more.

E(tricks) = trumps + (HCP-20)/3 + SST + e

Flat pattern pairs reduce tricks. It helps if one partner holds a singleton.


With a =5+5 fit bid game with 19 HCP, provided one partner holds a singleton.