| |
BGonline.org Forums
Non-doublet Replies: Five rules for beginners
Posted By: Axel Reichert In Response To: Non-doublet Replies: Four rules for beginners (nack ballard)
Date: Tuesday, 18 April 2017, at 6:57 p.m.
Hi Nack,
many thanks for your quick sketch, which set me onto the right track. Your proposal almost did the job to get below 0.08, but not quite: We get
32D-43 [D Z19 S91]
As you can see, 43Z avoids this blunder (as well as 52D-43 [D Z30 S64]) and (I checked this) does not create new ones. So this would be an easy change for the corresponding rule. There is only one other case with an error > 0.08:
54D-41 [$ S81]
This is covered (along with other similar blunders) by Jeremy's rule 4, "Slot Aces after [...] Down", my 8b. So we have (to get to 0.08 error or below):
1. Hit > 4
2. P, 65R
3. Hit 4, 3
4. If hit: Move backmost checker
5. 43Z, S (or most similar legal alternative)
5a. D-1x$
Then I lowered the threshold to 0.06. I found
43D-64 [R S01 P71]
which is covered by Jeremy's rule 5B, "Run past Down" (what a great formulation), my 5a. Also
64S-65 [K D75]
which is covered by Jeremy's rule 1Bii, "After 6x split, 65 hits on Bar and Ace", my 4a. And then there are a couple of cases where (after a hit) moving the rear-most checker is not best, e.g.
41$-64 [R U64]
These are covered by Jeremy's rules 1B and 3A (my 4 and 7). That's it, the 32Z-64P you mentioned loses only 0.055 and is thus not as bad as you thought. With these few additions, we have a mixture of your proposal and mine, a rule set to play non-doublet replies without blunders (< 0.06):
1. Hit > 4
2. P, 65R
2a. 64R past D
2b. 6xS-65K
3. Hit 4, 3
4. If hit from mid: S (or D). If hit from non-mid: D (or S)
5. 43Z, S (or most similar legal alternative)
5a. D-1x$
With a rule set that short, I can now afford to play around with the exceptions needed to get down to 0.04 errors. Thanks again for your help!
Axel
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.