| |
BGonline.org Forums
Rollout
Posted By: Stick In Response To: Rollout (Chuck Bower)
Date: Thursday, 5 February 2009, at 6:02 a.m.
Truth be told, I wasn't going to answer this until you put forth that I ought to be answering it. I briefly considered it this morning after he posted it but didn't think it was worth the time defending either my stance or the rollout results. I'm not even lit up, sorry to disappoint.
First off I was wrong that nobody voted 11/6 as Neil Robins pointed out. I didn't strictly count the votes because there was no need to, I knew how everyone voted and I didn't remember anyone voted 11/6, my most noble of apologies.
Secondly, on to "That's probably because 11/6 is wrong. One should never quit thinking and just trust the bot blindly"
I would hope that everyone knows I don't blindly accept rollout results (as Phil Simborg could tell you when I harass him for blindly accepting evaluations). However, when I have no reason to doubt the bot's ability to play out a position I'm much more likely to accept a 6k rollout from GNU than any one liner from any person.
This rollout meets that criteria. There's no reason in my mind I would doubt GNU making the correct decisions in upcoming plays and cube decisions, esp. compared to any human.
I didn't look hard into the position when posting it, I merely noted the 4 was the only forced part of the move and I didn't know where I would play the ace. This is all the further I need to look and then I let y'all vote on it. If I was to look further here's what may have taken place.
22/21 gives me 5s & 6s to escape. [66 65 64 63 62 61 55 54 53 52 51] This should total 20 as all double directs do. However, I can see that 55 will not likely be used to escape as it's such a powerful offensive number filling in my board so nicely bringing that number down to 19 escaping numbers. I also note that some of these escaping numbers may be duplicated if my opponent enters as pointing numbers though I wouldn't put that into my mental calculation, only a note.
11/6 gives me 6s to escape along with 51 for a total of 13 escaping numbers compared to 19 for 22/21. If that was the only thing I was worried about the choice would be clear. As is often the case, it's not the only consideration.
22/21 is both a pro and a con. Moving up to the edge of a prime preparing to escape is often good but here the +/- of it is a bit muddled. First off, not that the 2pt is made so the non entering 2s now play good, 6/4*, whereas before they would have just played, 13/11. A small notch in favor of 11/6.
It's also clear that 11/6 will be the play that wins the most gammons. It now adds some key small pointing numbers (that can't be used to escape). This is a slightly larger + in my opinion arguing for 11/6. Also, whenever you have no spares on the 6pt, a spare on the 7pt, and an ace to play, you'd better damn well be sure the other ace is more productive. 7/6 is such a natural play for that ace.
As far as gammons lost are concerned I would have considered both plays to be a push, gammons won favor 11/6, and games won favor stepping up. In guessing the trade off I would have guessed that the normal 2 to 1 trade off needed for money was sufficiently met by 11/6.
Now I've just written a lot of bullshit and I'm still unsure what the best play is, all I'm sure of now is that I'm not sure, which is where the rollout comes into play and reaffirms what my beliefs may have been if I had analyzed pre rollout. I completely trust this rollout and am no way convinced that stepping up is better than continuing to the 6pt.
Stick
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.