As we have three problems, this post will again be a long one. Next post will be shorter, I promise.
Exercise 1:
In the position below give yourself any roll containing 2 or 4 - numbers which could hit his checker on 4pt.
Except for 11, 22 and 42, all other rolls could hit but you cannot make a point, so if you hit on 4pt, you're leaving a blot he could hit from the bar and send your third checker back.
Would you hit in those positions? In all, in none, in some of them (which one: 21, 41, 32, 52, 62, 43, 54)?
 is Player 2score: 0 pip: 162 | ||||||||||||||||
 | Unlimited Game | |||||||||||||||
| pip: 167 score: 0  is Player 1 | ||||||||||||||||
| XGID=-a--a-E-C---dEa--c-e----B-:0:0:1:00:0:0:0:0:10 | ||||||||||||||||
| to play 21; 41; 32; 52; 62; 43; 54 | ||||||||||||||||
I will repeat the important rule from the previous post:
Once your opponent has split to his 21pt or 20pt, don't let him make his advanced anchor! Attack the blot on 4pt or 5pt!
Maybe it is contra-intuitive for some of the rolls in above diagram, but hitting on 4pt is correct in all given positions!
As rollouts for all non-trivial second roll positions are already done, I will just post a links to the rollouts on bgonline.org site:
21: 24/23 6/4*
41: 24/23 8/4*
32: 13/10 6/4*
52: 13/8 6/4*
62: 24/18 6/4* or 13/7 6/4*
43: 13/10 8/4*
54: 13/4*
Couple of short comments:
1. With 32 and 43 rolls, if your opponent is on 4pt then it is better to keep your blot on 4pt and bring another builder on 10pt that protects the blot on 4pt. If your opponent were on 3pt, then hitting twice would have been correct.
This is also true if your opponent is on 5pt and you have 41 to play. It is better to hit and stay on 5pt, rather than hit twice.
You should hit twice if the first hit is on a lower point (2pt or 3pt), but hit and stay if it's on a higher point (4pt or 5pt).
2. Hitting is correct even if you hit from 8pt. That's how important it is to prevent your opponent from making advanced anchor.
Exercise 2.1:
| is Player 2 score: 0 pip: 166 | ||||||||||||||||
| | Unlimited Game | |||||||||||||||
| pip: 151 score: 0 is Player 1 | ||||||||||||||||
| XGID=aaB-B-C-A---dEa--c-e----B-:0:0:1:00:0:0:0:0:10 | ||||||||||||||||
| on roll, cube action? | ||||||||||||||||
Would you double in this position? Is this take or pass? How sure are you in doubling and taking decisions?
This is only your second roll (opponent started with 32: 24/21 13/11, you played 44: 8/4*(2) 6/2(2), opponent danced, your turn to play), but you have cube decision already.
It is important not to throw dice or click on Roll dice automatically (or even worse: have Auto roll turned on). One of the famous backgammon rules is:
Every roll is a new cube decision.
If you keep that in mind, it will improve your game more than you think.
Look at our position. You have three home board points, you are ahead in the race, your opponent is on the bar and has a blot on your 14pt, he has no new home board points.
Also, an important consideration are the market losers.
Even if you roll some below average number so you don't close a new home board point nor hit a new blot, and your opponent dances again - you have lost your market - even with your worst roll (such as 21). With your three home board points, he dances 25% of a time (9/36 rolls).
On the other hand, if you close another home board point, you'll have four points closed and he will dance on almost half of his rolls (16/36).
Your stronger position, market losers and gammon potential (almost half of your wins will be gammons) make this position a strong double.
If your opponent anchors, he has a playable position and he can win from this position often enough to make this an easy take.
At this page I tried to explain how to read analysis numbers from backgammon software.
Here is the rollout result for cube decision:
| is Player 2 score: 0 pip: 166 | ||||||||||||||||
| | Unlimited Game | |||||||||||||||
| pip: 151 score: 0 is Player 1 | ||||||||||||||||
| XGID=aaB-B-C-A---dEa--c-e----B-:0:0:1:00:0:0:0:0:10 | ||||||||||||||||
| on roll, cube action? | ||||||||||||||||
| Analyzed in Rollout | No double | Double/Take |
| Player Winning Chances: | 64,26% (G:30,09% B:1,77%) | 64,47% (G:30,49% B:1,84%) |
| Opponent Winning Chances: | 35,74% (G:8,28% B:0,38%) | 35,53% (G:8,27% B:0,38%) |
| Cubeless Equities | +0,517 | +1,052 |
| Cubeful Equities | ||
| No double: | +0,703 (-0,067) | ±0,007 (+0,696..+0,711) |
| Double/Take: | +0,771 | ±0,010 (+0,761..+0,781) |
| Double/Pass: | +1,000 (+0,229) | |
| Best Cube action: Double / Take | ||
| Rollout details | ||
| 5184 Games rolled with Variance Reduction. Dice Seed: 45297676 Moves and cube decisions: 3-ply | ||
| Double Decision confidence: | 100,0% | |
| Take Decision confidence: | 100,0% | |
| Duration: 49 minutes 33 seconds | ||
eXtreme Gammon Version: 2.00.161.beta
Exercise 2.2:
Let's say you doubled, he took the cube. We reach this position.
You are now in a blitz mode and your game plan is to keep your opponent on the bar as long as possible.
What would you play with: 33, 44, 65?
| is Player 2 score: 0 pip: 166 | ||||||||||||||||
 | ||||||||||||||||
| Unlimited Game | ||||||||||||||||
| pip: 151 score: 0 is Player 1 | ||||||||||||||||
| XGID=aaB-B-C-A---dEa--c-e----B-:1:-1:1:00:0:0:0:0:10 | ||||||||||||||||
| to play 33, 44, 65 | ||||||||||||||||
You can keep your opponent on the bar either by hitting or closing new home board points.
With 33, you should switch points and play 4/1*(2), putting his second checker on the bar against your three point board.
He has only 25% chance to enter both checkers. On the other 75% of a time he will either stay on the bar with both checkers (9 rolls) or enter with only one checker (18 rolls).
Top five moves for 33:
1. | XG Roller+ | 13/10(2) 4/1*(2) | eq: +0,591 | |||
| ||||||
| 2. | XG Roller+ | 24/21 13/10 4/1*(2) | eq: +0,542 (-0,049) | |||
| ||||||
| 3. | XG Roller+ | 13/7 4/1*(2) | eq: +0,483 (-0,109) | |||
| ||||||
| 4. | XG Roller+ | 24/21(2) 4/1*(2) | eq: +0,443 (-0,149) | |||
| ||||||
| 5. | XG Roller+ | 24/21(2) 13/10(2) | eq: +0,414 (-0,177) | |||
| ||||||
eXtreme Gammon Version: 2.00.161.beta
With 44, you should make your 5pt by playing 13/5(2).
This is much more important than making your advanced anchor, as your game plan should be to blitz your opponent and to keep him on the bar as long as possible. He's not threatening your back checkers.
Top five moves for 44:
| 1. | XG Roller+ | 13/5(2) | eq: +0,717 | |||
| ||||||
| 2. | XG Roller+ | 24/20(2) 13/9(2) | eq: +0,594 (-0,123) | |||
| ||||||
| 3. | XG Roller+ | 13/9 13/1* | eq: +0,555 (-0,162) | |||
| ||||||
| 4. | XG Roller+ | 24/20 13/1* | eq: +0,537 (-0,180) | |||
| ||||||
| 5. | XG Roller+ | 24/20(2) 13/9 8/4 | eq: +0,491 (-0,227) | |||
| ||||||
eXtreme Gammon Version: 2.00.161.beta
With 65 you cannot safely hit nor make a new home board point.
Playing the obvious 24/13 is a big error as it puts sixth checker on the midpoint (13pt), not addressing the main problems of this position.
Much better is covering 8pt blot and splitting to 18pt with 24/18 13/8.
But even better is to keep your initiative and put another checker on the bar with 13/7 6/1*. You unstack the midpoint, bring another checker in the attack zone and put second checker on the bar. All this raises your gammon probability.
Top five moves for 65:
| 1. | Rollout1 | 13/7 6/1* | eq: +0,366 | |||
| ||||||
| 2. | Rollout1 | 24/18 13/8 | eq: +0,330 (-0,035) | |||
| ||||||
| 3. | XG Roller+ | 24/18 6/1* | eq: +0,290 (-0,076) | |||
| ||||||
| 4. | XG Roller+ | 24/13 | eq: +0,288 (-0,078) | |||
| ||||||
| 5. | XG Roller+ | 13/8 13/7 | eq: +0,287 (-0,078) | |||
| ||||||
| 1 2592 Games rolled with Variance Reduction. Dice Seed: 45297676 Moves and cube decisions: 3-ply | ||||||
eXtreme Gammon Version: 2.00.161.beta
For the next time, what would you play here and why:
| is Player 2 score: 0 pip: 147 | ||||||||||||||||
  | | Unlimited Game | ||||||||||||||
| pip: 163 score: 0 is Player 1 | ||||||||||||||||
| XGID=-b---BD-B---cE---c-e--b-B-:0:0:1:54:0:0:0:0:10 | ||||||||||||||||
| to play 54 | ||||||||||||||||
I would split my back checkers and bring a builder down to the 8. One of my goals right now in the game is to build an advanced anchor, but I also want to continue building my home. So, my choice is 24/20, 13/8
ReplyDeleteAnother option is 13/8, 13/9, but this is less good because I want to split my back checkers before he build another point is home board.
Running 24/15 isn't really a good idea because we're slightly behind in the pip count and we don't want to minimise contact yet.
other moves (8/3, 8/4 OR 8/3, 6/2) are just stupid. of the remaining moves, 13/4 is less stupid, but does nothing to further our goals of making an anchor and it has too much risk with too little reward.
OK, big blunder for me on the 33 play (that's why I feel allowed to post here, I'm still a "backginner").
ReplyDeleteHmm... this looks like a 31P-55P-54 (for those who don't understand, I'm using Nactation).
Being behind in the race and having the 5pt made, I would try to prime the opponent with 13/9 13/8.
With the 3pt made and spares on the 6pt and 8pt (10 total checkers in the zone), Black wants to blitz. 24/20 would be then too dangerous.
We are not in a hurry to anchor because Black's back checkers are still on our ace point. If Black had already split, we should hurry for an anchor, but now I don't see the necessity to do it.
13/4 and 24/20 13/8 seem to be the only reasonable alternative candidates, because they provoke contact.
I split with 24/20 13/8. I think there are some other viable plays - running with 24/15, bringing down two builders with 13/8 13/10 - but I like the split best. With black having made a homeboard point I need to get off the ace point before I'm stuck behind a blockade, the midpoint is crying out to be unstacked, and my stripped 8 point could really use a spare.
ReplyDeleteI like to split early - as the game goes on, it becomes harder to get off the ace point and harder to make the decision of when it's time to abandon the ace point. By splitting early, I avoid difficult and costly decisions later - yes, it's sometimes a small error, but I'd rather make a small error now than a big one later.
@Turz - everybody is allowed to post here :) I hope that even intermediate players will learn something from these posts.
ReplyDeleteI think many players would forget to hit with 33, because we know that 4pt is more important than 1pt.
But in blitz - every home board point is of the same importance. And by switching to 1pt, we have more builders for our empty points and more chance for a successful blitz.
I would play 13/8 13/9. I consider the blot on the 9pt to be a useful builder for my next roll. Hopefully I don't get hit! :)
ReplyDelete13/8 13/9.
ReplyDeleteI need to make a prime right now. And with spares on 6 and 8 points and the single one on 9, I can hit if opp splits.
24/20 seems to be not worth the risk. I put no pressure on my opp so he is likely to hit me next time.
a intersting point would be: everyone should add how sure he or she is about the decision from 1 - 5 (5 very sure, 1 very unsure...)
ReplyDeleteAs far as I remember Stick states somewhere if your opponent makes an inner point with a double don't touch your back checkers. 13/8 13/9 here for me. If the three point was made by a 53 it would split with 13/8 24/20. After 53 black has to give up his eight point to hit from there and two checkers less in the attacking zone.
ReplyDeleteI would go for pthalos 13/8 and 24/20
ReplyDeletetraxxy