| |
BGonline.org Forums
Doubling with just 22.3% chance of winning
Posted By: Michael Sullivan In Response To: Doubling with just 22.3% chance of winning (bb766)
Date: Friday, 19 July 2013, at 3:11 a.m.
I will answer your second question, since Bob already got the first.
Think about what else could have happened in those previous rolls. Gnu could have pulled ahead further than you are now -- would that have turned the position into a pass? How much better could gnu get before you would be tempted to pass this cube? Well it would have to at least be a solid favorite, since you'd prefer being 50-50 with the match on the line to being down 4-2, right?
So, the question is, how often does gnu lose it's market, versus how often does it end up doing really badly and having no chance at all?
If on previous rolls, a 6-6 followed by a 2-1 from you would still have left you the favorite, then why should gnu double? In the position you post here, any double 2-2 or higher by gnu followed by a non-double (or 1-1) from you will lose the market by a mile, so this turns out to be almost like a last roll position. In a previous position, even though gnu had a higher overall chance of winning, it's likely that a high double followed by a bad roll from you would merely have brought gnu to a good position with a small-moderate market loss, and a small/medium double might even leave you with a take.
It is never correct to double (except for automatic doubles) when you do not have a market loser, and it is usually correct to hold the cube when you have only small market losers (your best sequences result in a pass that is fairly close).
The closer you get to a last roll, the less likely you have only small market losers, even if your position is worse.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.