opening 2-1 / 4-1
Dave Slayton
of the good recent books on the subject. I was weaned originally on books
like The Cruelest Game by Barclay Cooke. I learned that with an opening 2-1
or 4-1, you bring a man down and slot the 5-point. Then I was reading the
terrific new book by Kit Woolsey and Hal Heinrich called "New Ideas in
Backgammon" where they say on page 3 "Modern theory has resurrected the
early split--for example, an opening roll of 4-1 is now usually played 13/9
and 24/23..." and I see that Jellyfish does this (at least with 2-1, I
assume with 4-1). I thought wow, that's interesting, I didn't expect that.
Then I'm reading Robertie's equally new book "501 Essential Backgammon
Problems" and the very first problem is how to play 2-1 as an opening roll,
and Robertie slots the 5-point.
So, what IS the modern theory? I guess there is still plenty of
disagreement about which play is "best", if different experts writing
different books both published this year can't agree on which play is
"best".
Opinions?
dmg
The splitting versus slotting plays you describe are extremely close.
I tend to give a little more credence to the solutions in the
Woolsey/Heinrich book than to Robertie's because theirs are based on
very extensive rollouts. (I assume Robertie has taken rollout data
into account as well, though I've never heard him make reference to
it.) Also, I've examined rollout results from many different
sources--done using different versions of different bots, set at
different levels, over several years--and the splitting plays have
consistently come out slightly ahead of the slotting plays.
I'm in the process of rolling out all the plays in Robertie's "501
Essential Backgammon Problems," and my results for Position 1,
indicate that splitting with an opening 2-1 is 0.016 better than
slotting. So the trend continues.
However, there seems to be a general consensus that bots don't play
backgames at quite the same level as they do other types of positions.
This may mean that slotting plays, which tend to lead to more
backgames than splitting plays, might be a bit undervalued by bots in
both their evaluations and rollouts (since the bots would tend to
steer toward other types of games). Still, each new generation of bots
seems to understand backgames, and the timing considerations involved,
better than the one before, and the splitting plays continue to come
out ahead in the rollouts.
Ultimately, splitting and slotting seem so close that I think the
choice comes down to personal style and preference; we are better off
playing what feels comfortable to us. Most of the top players I watch
seem to play the splitting plays if they are even or ahead in the
match and switch to the slotting plays if they are trailing or needing
a gammon.
Dean
("Chase" on FIBS and GamesGrid)
_______
To respond via email, replace "USERNAME" with "demiga" in my address.
e...@home.com
>
> I am coming back to backgammon after a long drought, and have acquired some
> of the good recent books on the subject. I was weaned originally on books
> like The Cruelest Game by Barclay Cooke. I learned that with an opening 2-1
> or 4-1, you bring a man down and slot the 5-point. Then I was reading the
> terrific new book by Kit Woolsey and Hal Heinrich called "New Ideas in
> Backgammon" where they say on page 3 "Modern theory has resurrected the
> early split--for example, an opening roll of 4-1 is now usually played 13/9
> and 24/23..." and I see that Jellyfish does this (at least with 2-1, I
> assume with 4-1). I thought wow, that's interesting, I didn't expect that.
>
> Then I'm reading Robertie's equally new book "501 Essential Backgammon
> Problems" and the very first problem is how to play 2-1 as an opening roll,
> and Robertie slots the 5-point.
>
> So, what IS the modern theory? I guess there is still plenty of
> disagreement about which play is "best", if different experts writing
> different books both published this year can't agree on which play is
> "best".
Eerie. I'm in exactly the same situation, having just returned after
a long long time, and also having started with Cooke's book, and also
wondering the exact same thing, about splitting vs. slotting. I was
going to compose a post for r.g.b this weekend, asking exactly the
same question. Thanks.
Gregg Cattanach
After doing some EXTENSIVE rollouts of the 2-1 and 4-1 opening moves, the
net equity (for money) for the various moves is:
24/23 13/11: .001
13/11 6/5 : -.011
24/23 13/9 : -.002
24/20 24/23: -.023
13/8 : -.023
13/9 6/5 : -.033
With 2-1, the slot of the 5 point is still pretty popular, as well it should
be, but the split is technically just a bit better. When you need gammons
more than your opponent, then slotting is most likely best.
With 4-1, the old 'standard' play of 13/9 6/5 is actually worse than 13/8 !!
So most people are splitting with this move now. The problem with 13/9 6/5,
is there are several crushing numbers that really put you on the defensive
very quickly, 2-2, 4-4 and even 1-1 all put you back quite a way, as well as
3-5 and 6-2. There are too many ways for your opponent to seize the
initiative....
Just a little grist for the mill....
--
Gregg Cattanach
Zox_ at GamesGrid, Zone
http://gateway.to/backgammon
ICQ #2266410
gcattana...@prodigy.net
"Dave Slayton" <sla...@sprynet.com> wrote in message
news:8mc9e3$165$1...@slb7.atl.mindspring.net...
> I am coming back to backgammon after a long drought, and have acquired
some
> of the good recent books on the subject. I was weaned originally on books
> like The Cruelest Game by Barclay Cooke. I learned that with an opening
2-1
> or 4-1, you bring a man down and slot the 5-point. Then I was reading the
> terrific new book by Kit Woolsey and Hal Heinrich called "New Ideas in
> Backgammon" where they say on page 3 "Modern theory has resurrected the
> early split--for example, an opening roll of 4-1 is now usually played
13/9
> and 24/23..." and I see that Jellyfish does this (at least with 2-1, I
> assume with 4-1). I thought wow, that's interesting, I didn't expect
that.
>
> Then I'm reading Robertie's equally new book "501 Essential Backgammon
> Problems" and the very first problem is how to play 2-1 as an opening
roll,
> and Robertie slots the 5-point.
>
> So, what IS the modern theory? I guess there is still plenty of
> disagreement about which play is "best", if different experts writing
> different books both published this year can't agree on which play is
> "best".
>
> Opinions?
>
>
>
>
Michael Manolios
rollouts showed that:
You should always split with 41. 24/20 24/23 is also acceptable when
gammons are of no use to you if I remember well (e.g. you are ahead in
a post - Crawford game).
With 21 you should:
a) Split when playing for money, or the score is irrelevant (e.g. 0-0
in a 7 point match), or gammons don't count for you (e.g. you are ahead
in a post - Crawford game).
b) Slot when in a double match point, or gammons don't count for your
opponent (e.g. you trail in a post - Crawford game).
In article <8mc9e3$165$1...@slb7.atl.mindspring.net>,
--
Michael Manolios (mann on FIBS, Glass on GG)
We play one and only money game session through our whole life...
Sent via Deja.com http://www.deja.com/
Before you buy.
e...@home.com
>
> After doing some EXTENSIVE rollouts of the 2-1 and 4-1 opening moves, the
> net equity (for money) for the various moves is:
Thanks. This is just the sort of analysis I assume is available now,
thnks to computers, that wasn't 25 years ago. Is this type of thing
available in current literature, and if so, what do people recommend?
Dave Slayton
split is indicated too, since you were forced to play the 5 from the
midpoint, and so you don't even have an extra builder to try to cover the
5-point with if you slot. But I wonder what the equity would be for playing
24/18...if he hits but doesn't cover, you get lots of return shots. 6-6 or
1-1 would be unwelcome by him, but not much more so than if you'd made any
other play.
And then there's 6-2, with which I used to slot the 5-point. So what's the
suggested play there if you don't do that? 24/18,13/11? 24/16?
Then there's 3-4, about which Peter Backgren has started a new note string,
and advocates 24/21,13/9, which seems to have a lot of merit.
Splitting to the 22-point with a 2 invites 5-5 from your opponent, while it
seems one of the reasons to split with 4-1 is to avoid the crushing 4-4
return shot if you slot. So apparently the split to the 22-point is
ill-advised for the same reason?
But splitting to the 21-point doesn't seem so bad, putting a blot on a
well-advanced point that's not as tasty for your opponent as the 20-point
where he'll be dying to hit you.
So this brings into question the rolls 2-3, 3-4, 3-5, and 3-6. What sort of
equity would you get, I wonder, splitting with the 3 in these cases, versus
the other reasonable plays for these numbers?
Robert-Jan Veldhuizen
wrote:
>This makes me wonder about some other rolls, like 5-1. I'd guess there the
>split is indicated too, since you were forced to play the 5 from the
>midpoint, and so you don't even have an extra builder to try to cover the
>5-point with if you slot.
Yes, 24/23 13/8 is a bit better according to the bots than 13/8 6/5 for
the reasons you state.
>But I wonder what the equity would be for playing
>24/18...if he hits but doesn't cover, you get lots of return shots. 6-6 or
>1-1 would be unwelcome by him, but not much more so than if you'd made any
>other play.
If I remember correctly 24/18 is a serious contender and better than the
slot (at least if gammons count), it just doesn't seem popular at all,
probably because it's a bit of a boring move.
>And then there's 6-2, with which I used to slot the 5-point.
I think the bots definitely don't like 13/5 anymore, probably because
the risk of the slot isn't compensated for by developing the stack on
your sixpoint with this move.
>So what's the
>suggested play there if you don't do that? 24/18,13/11? 24/16?
24/18 13/11 is common, 24/16 about equal or maybe even a bit better,
especially when gammons don't count for you but they do count for your
opponent.
>Then there's 3-4, about which Peter Backgren has started a new note string,
>and advocates 24/21,13/9, which seems to have a lot of merit.
I think this move and 24/20 13/10 are very close. Peter's arguments for
splitting to the four are right, but with splitting to the five you have
a better chance of making this anchor, and the fivepoint anchor is
definitely worth more than the fourpoint anchor. Also, a builder on 10
is perhaps worth just a tiny bit more than one on 9. It seems like the
extra risk just about equals the extra gain.
13/10 13/9 is slightly worse, but still atrractive if you want to gammon
your opponent.
>Splitting to the 22-point with a 2 invites 5-5 from your opponent, while it
>seems one of the reasons to split with 4-1 is to avoid the crushing 4-4
>return shot if you slot. So apparently the split to the 22-point is
>ill-advised for the same reason?
No, not according to the bots. The split to 22 only risks 5-5, whereas
building and slotting with 4-1 risks getting checkers send back, also
with 2-2 and 6-4, that's four times as often.
Sure 5-5 is bad, but the bots still like to split to 22 with a 5-2.
13/11 13/8 is about equal though if I remember correctly. The matchscore
can be a deciding factor here.
>But splitting to the 21-point doesn't seem so bad, putting a blot on a
>well-advanced point that's not as tasty for your opponent as the 20-point
>where he'll be dying to hit you.
>
>So this brings into question the rolls 2-3, 3-4, 3-5, and 3-6. What sort of
>equity would you get, I wonder, splitting with the 3 in these cases, versus
>the other reasonable plays for these numbers?
24/21 13/11 is slightly favored above 13/10 13/11, the latter is good if
you need gammons more than your opponent though.
3-4: see above.
3-5? I don't think there's any question about that move! Make the
innerboard point with 8/3 6/3. A typo perhaps?
3-6: The split to the barpoint is just about equal to running 24/15. The
mattchscore can be a deciding factor here.
--
Robert-Jan/Zorba
Alex Zamanian
Robert-Jan Veldhuizen wrote:
>
> 3-5? I don't think there's any question about that move! Make the
> innerboard point with 8/3 6/3. A typo perhaps?
>
According Snowie, splitting with a 5-3 (13/8 24/21) is quite reasonable
with the opening roll. I did a large cubeful rollout with Snowie 2
(checker play 1-ply, cube handling 1-ply), and splitting came out only
about .015 behind making the 3 point.
It might be a good move to try when you don't need a gammon, and when
gammons count against you a lot, like when you're winnning 1-away 2-away
Crawford.
I think a weak player should play this move against a stronger player
since the holding games that splitting leads to are easier to play than
the blitzing and priming games that making the 3 point leads to.
-Alex
Dave Slayton
>If I remember correctly 24/18 is a serious contender and better than the
>slot (at least if gammons count), it just doesn't seem popular at all,
>probably because it's a bit of a boring move.
it's more boring than 24/23,13/8?? not for me..... :)
dmg
<rjv...@iname.com> wrote:
|On Fri, 4 Aug 2000 08:38:55 -0600, "Dave Slayton" <sla...@sprynet.com>
|wrote:
|>Then there's 3-4, about which Peter Backgren has started a new note string,
|>and advocates 24/21,13/9, which seems to have a lot of merit.
|
|I think this move and 24/20 13/10 are very close. Peter's arguments for
|splitting to the four are right, but with splitting to the five you have
|a better chance of making this anchor, and the fivepoint anchor is
|definitely worth more than the fourpoint anchor. Also, a builder on 10
|is perhaps worth just a tiny bit more than one on 9. It seems like the
|extra risk just about equals the extra gain.
Perhaps there is something I'm not considering, but I've always felt
that a builder on the 9-point is a bit stronger than one on the
10-point, and I've assumed this is one reason why 24/21 13/9 does so
well in the rollouts.
I prefer the 9-point for several reasons:
1) It bears on more inner board points, giving me a good chance to
build a strong inner board, a prime, or to carry out a successful
blitz.
2) I'd rather make the 9-point than the 10-point. A prime from the
4-point to the 9-point is more valuable and easier to build and manage
than one from the 5-point to the 10-point.
3) Many of the point-making numbers that result from having a checker
on the 10-point duplicate those that already play well from the 6- and
8-points. By placing a checker on the 9-point instead, rolls such as
41 or 52 become very good.
Of course the down side is that a blot on the 9-point is hit more
often, either from the 24-point or by return shots. I also like going
after the best anchor available on the defensive end. For these
reasons I most often play 24/20 13/10 with an opening 43, playing
24/21 13/9 occasionally for variety.
Robert-Jan Veldhuizen
wrote:
>(speaking about the play of 5-1)
Hehehe... well perhaps it's more that it feels like you rolled 6-0 if
you play 24/18 with a 5-1 :-)
--
Robert-Jan/Zorba
Robert-Jan Veldhuizen
>On Fri, 04 Aug 2000 18:44:18 +0200, Robert-Jan Veldhuizen
><rjv...@iname.com> wrote:
>|Also, a builder on 10
>|is perhaps worth just a tiny bit more than one on 9. It seems like the
>|extra risk just about equals the extra gain.
>
>Perhaps there is something I'm not considering, but I've always felt
>that a builder on the 9-point is a bit stronger than one on the
>10-point, and I've assumed this is one reason why 24/21 13/9 does so
>well in the rollouts.
In itself, a builder on 9 is worth more than one on 10 for the reasons
you gave. What I meant to write was that *bringing down* a builder to 10
might be just a tiny bit better than bringing one down to 9.
What's the difference? Well, a builder on 10 can still become a builder
on 9 later with a one; and a builder on 10 can also be used to build the
9 point, but never the other way around.
During the openings phase it might be worth keeping just this tiny bit
of extra flexibility. Personally, I like bringing down a builder with a
two to the 11 point best of all, because of this. It gives you the most
options for the follow-up. For the same reason, I think getting a
builder to 10 is a bit better than bringing one down to 9 on the opening
roll. Of course the number of hits also plays a role in this.
BTW, I disagree slightly about the prime. The further back a prime is,
the more useful it is, very generally speaking (assuming it's a mutual
priming game), as this gives you better timing to roll it forward and
thus win a priming battle. If your opponent has a more advanced anchor
or is threatening to get one, it's a different story of course.
A prime from 5 to 10 is definitely more difficult to *create* than one
from 4 to 9 and as such less common, but I think this is actually a
reason why you'd try to keep the checkers in the outfield as far back as
possible: to give you most options. Trying to make the sixth point of a
prime on ten will generally be harder than if the sixth point to make is
the nine point.
But all this is of course first-class nit-picking or whatever you'd call
it... if there's any difference at all, it'll be immeasurably small I
suppose. Still always interesting to discuss things a bit. :-)
--
Robert-Jan/Zorba
Donald Kahn
>Dave Slayton wrote:
>>
>> I am coming back to backgammon after a long drought, and have acquired some
>> of the good recent books on the subject. I was weaned originally on books
>> like The Cruelest Game by Barclay Cooke. I learned that with an opening 2-1
>> or 4-1, you bring a man down and slot the 5-point. Then I was reading the
>> terrific new book by Kit Woolsey and Hal Heinrich called "New Ideas in
>> Backgammon" where they say on page 3 "Modern theory has resurrected the
>> early split--for example, an opening roll of 4-1 is now usually played 13/9
>> and 24/23..." and I see that Jellyfish does this (at least with 2-1, I
>> assume with 4-1). I thought wow, that's interesting, I didn't expect that.
>>
>> Then I'm reading Robertie's equally new book "501 Essential Backgammon
>> Problems" and the very first problem is how to play 2-1 as an opening roll,
>> and Robertie slots the 5-point.
>>
>> So, what IS the modern theory? I guess there is still plenty of
>> disagreement about which play is "best", if different experts writing
>> different books both published this year can't agree on which play is
>> "best".
>
>Eerie. I'm in exactly the same situation, having just returned after
>a long long time, and also having started with Cooke's book, and also
>wondering the exact same thing, about splitting vs. slotting. I was
>going to compose a post for r.g.b this weekend, asking exactly the
>same question. Thanks.
One piece of advice. Don't pay serious attention to Cooke's book. It
is charmingly written, but at least half the answers are long-way
wrong, according to present day expert play.
Most good players I know slot with 2/1. Of course if you don't get
hit you now have a major advantage. And if you get hit, and I have
not seen this specifically pointed out, you have a good six to play
from the bar, continuing the contest for your 5-point. Compare this
with the lack of a good six when your 4/1 slot is hit. Now your six
is an embarrassment. (In fact I see more slots with 5/1 than with
4/1.)
Some of us (M. Svobodny among others) have a little joke. We slot
with 4/1 but change to a split before picking up the dice. The joke
is that to slot with 4/1 is an insult to the opponent - he must be a
soft touch.
dk
e...@home.com
[...]
> One piece of advice. Don't pay serious attention to Cooke's book. It
> is charmingly written, but at least half the answers are long-way
> wrong, according to present day expert play.
OK. And why do the "experts" now think what they think? Because what they
think is in fashion and they're the experts, or because there's sound analysis
behind it. If it's because there's sound analysis, where is it? What books
can I find that present backgammon like the Cooke or Magriel or Jacoby books
but that are grounded in the types of analysis that have been presented in
responses to Dave Slayton's question.
Dave Slayton
school of backgammon was that splitting one's back men was viewed generally
as poor tactics by Barclay Cooke and others, cuz you would be leaving
yourself open to being blitzed, etc.
But I noticed from all this discussion and the various resources I was able
to find containing volumes of rollout data on opening rolls, that the modern
"best" play for ALL opening rolls except for the 4 that make good points
immediately (5-3,4-2,3-1,6-1) involves splitting the back men, and with the
added exception of 6-5, they all involve moving a single back man and a
single man from the midpoint.
(I say "best" cuz what's really really best appears to still be
controversial on some rolls like 5-2, 5-4, 3-2, and 4-3, where there are a
couple of possible good plays that come out very close in the various
rollouts, but regardless of that, the splitting plays for each are about as
good or slightly better than the other viable plays.)
So I find this splitting-is-evil vs. splitting-is-wonderful disparity
between the "old school" and the "new school" really interesting.
Robert-Jan Veldhuizen
wrote:
[...]
>So I find this splitting-is-evil vs. splitting-is-wonderful disparity
>between the "old school" and the "new school" really interesting.
This is just speculating from my side, but I think it has something to
do with psychology.
Neural nets just count wins and losses; they don't care how the game
went.
Human beings however, will probably remember games differently. It's not
much fun splitting with a 2 and then getting blitzed, with probably very
few if any unforced moves left in the game.
If there's an audience that'll tell you that's why you should never
split with a two, just after you got gammoned for 4 points like that,
it's hard to convince them (and after a while, yourself!) otherwise.
Also, if you play 4-1 like 13/9 6/5 and don't get hit, you'll probably
have lots of interesting plays coming up and your position will look
good and 'professional' to any watchers. Even if you lose such a game,
you'll probably feel like you've played a pretty good game and your
opponent just got incredibly lucky with that shot in the bear-off.
If one or both blots get hit, you'll still be facing an interesting
game, probably a backgame where you can really show your skill (ahum)!
If you just happen to roll that 6-6 behind your opponent's prime later
on, completely ruining your timing and effectively losing you the game,
it's easy to convince yourself that was just very bad luck; watchers
might agree since they were hoping for some spectacular endgame
probably. The fact that 6-6 AND 5-5 would've ruined your backgame for at
least 12 moves in a row for instance, giving a total probability of this
"horror" sequence to happen 1-(17/18)^12 =~ 50% of the games actually,
is easily forgotten!
The 'old school' didn't have much objective tools available when
compared to what we have now, so I think it's likely these psychologic
effects had an influence on the style of play.
BTW, I'm not saying some of the old school plays, especially the ones
that are very likely theoretically inferior according to nowadays bots,
are wrong because of this. Especially when playing weaker or stronger
opponents, or perhaps even opponents you know to have weak spots in
certain areas of the game, it might gain, perhaps quite a lot, to make
plays that are likely to lead to certain types of positions that you
think give you a relative advantage, like priming battles, backgames
(usually good when playing weaker opponents), races, mutual holding
games (usually good when playing stronger opponents).
--
Robert-Jan/Zorba
Don Hanlen
equity, which may be countered by one's style of play. The important
thing to remember is that where an optional roll occurs on the second
roll, one should consider what the opponent did. For example, if your
opponent moved 1-7 13-10 with a 6-3, it certainly changes your perspective
if you come back with a 4-3.
In article <8mp85h$gbt$1...@slb6.atl.mindspring.net>,
Dave Slayton <sla...@sprynet.com> wrote:
>One more comment on this issue of opening rolls: My impression from the old
>school of backgammon was that splitting one's back men was viewed generally
>as poor tactics by Barclay Cooke and others, cuz you would be leaving
>yourself open to being blitzed, etc.
>
>But I noticed from all this discussion and the various resources I was able
>to find containing volumes of rollout data on opening rolls, that the modern
>"best" play for ALL opening rolls except for the 4 that make good points
>immediately (5-3,4-2,3-1,6-1) involves splitting the back men, and with the
>added exception of 6-5, they all involve moving a single back man and a
>single man from the midpoint.
>
>(I say "best" cuz what's really really best appears to still be
>controversial on some rolls like 5-2, 5-4, 3-2, and 4-3, where there are a
>couple of possible good plays that come out very close in the various
>rollouts, but regardless of that, the splitting plays for each are about as
>good or slightly better than the other viable plays.)
>