It seems that some of you can download Rick Janowski's pdf on Take-Points in
Money Games and some of you can't.
Would a few of you have a go and let me know how you get on? It is quite
important that the file downloads OK else Rick's been wasting his time!
If you don't have Acrobat Reader it can be downloaded from the address
"michael.a.crane" <michael...@ntlworld.com> wrote in message
>Can I ask for some help here?
>It seems that some of you can download Rick Janowski's pdf on Take-Points in
>Money Games and some of you can't.
>Would a few of you have a go and let me know how you get on? It is quite
>important that the file downloads OK else Rick's been wasting his time!
I had IE 5.5 and Acrobat Reader 3.01 already installed, and it displays
fine through AR inside the IE window.
> Can I ask for some help here?
> It seems that some of you can download Rick Janowski's pdf on Take-Points in
> Money Games and some of you can't.
> Would a few of you have a go and let me know how you get on? It is quite
> important that the file downloads OK else Rick's been wasting his time!
All well and good here.
Be sure that folks downloading this pdf document are using a recent copy of
Acrobat. Some pdf documents created in Acrobat 4 cannot be read by 3.x readers.
BTW. Gave it a quick scan and my brain started to hurt. Any suggestions for this
>BTW. Gave it a quick scan and my brain started to hurt. Any suggestions for this
Not really I think. Here we have some very mathematical and very
theoretical stuff for sure %-)
Still there was something funny that I'd never thought about or read
about before and that might be an interesting puzzle.
Some us might have heard of the Kauder paradox; it's also in Robertie's
AB Vol.II. It exists because of the Jacoby Rule: (back)gammons only
count after a double. There are bg positions that are proper initial
doubles (activating (back)gammons), yet also proper beavers (after
taking, the opponent is the favorite with his cube ownership).
In Janowski's paper, I read about another paradox (I forgot the name),
also existing because of the Jacoby rule: bg positions that are proper
redoubles, but not initial doubles!
Normally, initial doubles can be offered earlier than redoubles, because
with a redouble you give your opponent cube access that he didn't have
before. So, a position that is an initial double might not be a
redouble; a redouble generally needs a bit more.
With the Jacoby rule however, an initial double activates gammons, which
a redouble does not. Simply put, if activating gammons in itself is
negative for the doubler, he might not want to give an initial double;
with gammons already activated he might want to redouble though.
In practice, it doesn't happen too often that you would like to double
when the gammon games are quite clearly *not* in your favour. Still it
doesn't seem at all impossible and it looks more natural to occur than
proper double/beaver positions.
So here's a nice puzzle, especially for the people that can rollout some
play with a bot: construct a position in a money game, Jacoby that is a
proper redouble yet not an initial double! :-)
Downloaded fine using RealDownload, look forward to working through it
Regards Richard Biddle
roll. Jacoby in use. Cube action?
| | X O O O O | | |
| | X O O | | |
| | O O | | |
| | | X | |
| | | | |
| | |BAR| |
| | | | |
| | | | |
| | | | |
| | X X X X X X | | |
| | X X X X X X | | |
| CubeValue: A. 1 centered, or B. 2 with X owning it
I think this position qualifies as a solution to your "Paradox". Here is a 1440 game JellyFish 3.0 Level 6 full-rollout, which it says its variance reduction makes equal to 32,000 games, and who are we mere humanoids to question this:
X wins = 67.1%
O wins 32.9%, of which 13.8% are singles, 15.2% are gammons, and 3.9% are backgammons.
First, the center-double case:
If he waits, X will hit 25/36 times and win with the cube; 11/36 times he will miss and be doubled out, since O has no reason to play on with gammons not activated:
25/36 x 1 + 11/36 x -1
= 69.44 - 30.56 = + 38.88 points in 100 games
If he doubles now, I'll use the win % JF gets even though in reality there'd be some small adjustment for O being able to turn the cube sometimes after being hit if X has an accident bringing this home after hitting.
67.1 x +2 + 13.8 x -2 + 15.2 x -4 + 3.9 x -6
= 134.2 - 27.6 - 60.8 - 23.4 = + 22.4 points in 100 games
It is apparent that if the cube is centered, X is better off to wait. Now, about that redouble:
If X holds the cube at 2, his wins will be the 69.44% we saw in the wait from the center case. This is higher by 2.34 % than his wins if the game is played out, and it should come entirely at the expense of O's single wins, not his gammons or backgammons, so I adjusted the numbers for the double-from-center by this amount:
2.34 x 2 = swing on this case in X's favor = 4.68 x 2 cube = 9.36 points in 100 games
22.4 + 9.36 = 31.76 points in 100 games
If instead X redoubles, he ends up with the same result as if he'd doubled from the center, except now he's on a 4 cube:
22.4 x 2 = 44.8 points in 100 games
So it would appear X should redouble here to get the higher equity, despite the gammon risk, even though because of Jacoby he could not double from the center with this position.