Join Bridge Winners
All comments by Han Peters
1 2 3 4 ... 17 18 19 20
You are ignoring the author of this comment. Click to temporarily show the comment.
I believe that some recent world champions play that a direct 3S is weak.
Jan. 15
You are ignoring the author of this comment. Click to temporarily show the comment.
Agree with Bob on both accounts.
Jan. 13
You are ignoring the author of this comment. Click to temporarily show the comment.
You can easily allow any style you like.
Jan. 12
You are ignoring the author of this comment. Click to temporarily show the comment.
Hi Csaba,

I think that what you sketch is very good with respect to information leakage, but the gain in comparison with traditional methods lies mostly in the rest of the play, not in the opening lead.
Jan. 12
You are ignoring the author of this comment. Click to temporarily show the comment.
Regarding the Qx doubleton, I didn't think about that. I didn't want to pick an extreme hand like AJxx xx KQxx xxx, where I think that bidding stayman would do much better, or xxxx AQ Kxxx Jxx, where nobody would be surprised if bidding Stayman would do poorly.

So, I picked one ace, one king, etc, and a hand that looked not so special to me. It seems like I could have done better.
Jan. 11
You are ignoring the author of this comment. Click to temporarily show the comment.
You would have to make a choice regarding selecting the lead, yes. But for pairs it seems reasonable to just take the leads that give the maximal number of tricks. I understand the flaw you mentioned, but it wouldn't bother me. :)
Jan. 11
You are ignoring the author of this comment. Click to temporarily show the comment.
I agree with you regarding the effect of putting the 4333's in 1NT-3NT. I wonder how large the difference will be though.

I found that when ending in 3NT after both stayman and puppet stayman, regular stayman won by about 0.07 imps. This was an earlier study where I didn't specify the north hand but gave a random 4-2-4-3 10-count. The number of hands used in different steps were also slightly different.

For the article I thought it would be nicer to work with a specific north hand.
Jan. 11
You are ignoring the author of this comment. Click to temporarily show the comment.
I don't think that it is obviously more useful at pairs. Admittedly, very few things are obvious to me.
Jan. 11
You are ignoring the author of this comment. Click to temporarily show the comment.
Meike said she would put the code in a github. I don't know what that means, but I think it is what Michael asked for.

When I have the time I will describe in this thread exactly which assumptions I made for those who are interested but prefer not to read the code.

There have been a number of very interested follow up suggestions. I'm not going to follow up on most of them, but I hope somebody else will.
Jan. 11
You are ignoring the author of this comment. Click to temporarily show the comment.
1. It seems to me that it is best to use the same evaluation for selecting leads as the evaluation that is used in later steps.

2. I tried to explain something using as little lingo as possible without saying incorrect things. While I know very little about statistics, I do think that your conclusion is incorrect.

3. I think you missed an important point. The question is not whether 0.49 is statistically significantly greater than 0. Common sense already tells us that. The goal is to get an accurate estimate.
Jan. 11
You are ignoring the author of this comment. Click to temporarily show the comment.
I don't know what you mean regarding the effectiveness of Puppet. I do think that puppet is less revealing, but the study was not able to confirm this.

Thanks for your other comment though.
Jan. 11
You are ignoring the author of this comment. Click to temporarily show the comment.
Ha, thanks Timo! I had this topic lying around for quite some time, but to be honest, a week ago I was still in doubt whether anybody would be interested at all. I am quite surprised by the reactions, but very glad to read them of course.
Jan. 11
You are ignoring the author of this comment. Click to temporarily show the comment.
In the Stayman auction 1NT - 2C - 2D - 3NT, it is not revealed that responder has spades. All puppet stayman auctions except 1NT - 3C - 3H do reveal this.
Jan. 11
You are ignoring the author of this comment. Click to temporarily show the comment.
Hi Charles,

Good points.

I don't think that it is worth writing a lot of code trying to make the study perfect. The numbers are inherently flawed because I use double dummy analysis. So instead of thinking “blasting loses 0,09 imps on average”, I prefer to think “the price of information is not negligible”. It is worth being in the wrong contract every once in a while if it gives the opponents less to go on.

And that is something to take with me to the bridge table.

By the way, my first name is Han.
Jan. 10
You are ignoring the author of this comment. Click to temporarily show the comment.
I agree Kit, it will take me some time but I will do it.
Jan. 10
You are ignoring the author of this comment. Click to temporarily show the comment.
Perhaps if we start calling it “Common Sense” instead of “Bayes' Theorem”, people would become less afraid of it.
Jan. 6
You are ignoring the author of this comment. Click to temporarily show the comment.
Erm, Roland…
Jan. 6
You are ignoring the author of this comment. Click to temporarily show the comment.
Comparing yourself to a waste producing corporation may not get you the sympathy you hope for.
Jan. 3
You are ignoring the author of this comment. Click to temporarily show the comment.
I think this rule is bizarre.
Jan. 3
You are ignoring the author of this comment. Click to temporarily show the comment.
It's “you're”, not “your”.
Jan. 2
1 2 3 4 ... 17 18 19 20
.

Bottom Home Top