Hello everyone! First of all: sorry if my post is not well formatted, it is my first time on this forum... I have read quite a few posts and really enjoyed the content and quality of the responses. So I have some questions about evaluating your winning chances in mid/late game, here is a spot that I got yesterday:

.. I was playing a 5 point match with a friend of mine and came up with an interesting spot.

The game was 1-1 and Villain had the cube (cube at 2). This position occured and he doubled, which is the right move (I am white , home board: lower right) :

http://imgur.com/cZRBzye

I correctly refused the game because I intuitively evaluated that I was winning more often the match at 1-3 than this specific game. Turned out my intuition was right since I had around 26% chances to win this match (according to XG). This 26% is lower than my 32% chances to win the match to 5 when Villain leads 3-1 (see equity table :http://www.dummies.com/how-to/conten...ble-handy.html).

Now, had the board been a little different (giving me over the 32 % chances to win as evaluated by the equity table which is based on the fact that Villain and I are of the equivalent strength), I would have needed to accept. For instance, XG evaluates my winning chances in the following position at around 33,3 % :

http://imgur.com/FsWnMre


So finally, the question: how do you guys evaluate your mid to late winning chances in this type of spot? I know that making little mistakes of a few % equity do not mean that much, but how do you guys estimate your winning chances in spots like these two positions? Do you have any guidelines ?

Huge thanks,
Jiffypoops

FYI: I also posted the situation on the reddit page. So for the curious ones, here are the responses on reddit (https://www.reddit.com/r/backgammon/..._midlate_game/)

Estimating winning chances in a position like this isn’t trivially easy, but it’s doable, and the more experience you have, the better you get at the process. I’ll go through what my thought process over the board would look like, and hopefully you’ll find this useful.

Before we get started, let me just point out that I don’t have any formula here that’s going to solve the problem. There are racing formulas and bearoff formulas that can solve most races and bearoffs, but positions with any significant contact will require more thought than that. The race will always play a role, but we need to look at all the other features of the position as well.

I always start evaluations in a match by figuring out what I would do for money, then adjusting for the match score. I like to tell people that money game play is the foundation of backgammon, and you can’t expect to play matches well unless you really understand how to handle the cube for money. So I’ll start off talking about the money game evaluation, then adjust for the match score at the end. Here are the questions I’d be asking myself:

(1) What kind of position is this?

I’d label this “Endgame – Bearing in against one man on bar.” It’s an endgame because there’s not much contact left, and bearing in against a man on the bar is a pretty common type of endgame, with its own set of rules and tips.

(2) What do I know about this type of game and what would I do if this were a money game?

The first thing I know is that the race is a big part of the evaluation, so let’s look at the race. In your diagram, White (you) has a count of 116, and Red (your opponent) is 110. His lead is a little less than 6%. If this were a race, our 8-9-12 rule says “No double and easy take”. But it’s not, so we continue.

Positions with one man back (on the bar or not) are worse than straight races. The contact favors the side coming home, because every turn he’ll have some rolls that point on you or pick and pass, and every turn that you spend on the bar is a disaster for you. Whatever your chances are, they’re worse than your chances in a straight race.

In “501 Essential BG Problems” I wrote about a few positions like this where your opponent’s side had a 5-point board. In that case, you actually had to lead in the race to take a double. Here you trail by six, so if your opponent had a 5-point board, the cube action would be double-pass.

So now we’ve got a couple of position types that bracket our actual position. If your position were a little better (a straight race) the cube action would be no double and take. If his position was a little better (a 5-point board) the cube action would be double and pass. That gives us some confidence that this position with a 4-point board is probably double and take.

Let’s also note that your opponent shouldn’t wait on the double. If he just throws an average sort of number (say 6-1, playing 14/7) and you dance again, he’s up 13 in the race and you’re still on the bar. I’d make that a solid pass, so he’s got plenty of market-losing sequences coming up. That implies doubling now, not waiting.

So our verdict for money is this: solid double and take, but only one roll away from double and pass. If I had to put a number on your winning chances over the board, I’d be guessing 25% to 30%, and more likely at the low end of that range.

(3) What do I do at this match score?

Translating a money game decision into a match play decision is sometimes tough, but with this particular position it’s the easiest part of the problem.

If you drop, your opponent takes a 3-1 lead in a 5-point match, which we generally refer to as ‘2-away 4-away’. If you know your match equity tables, your winning chances are in the 32% to 34% range, depending on whose table you use. The reason they’re higher than you might expect is that the leader’s cube is severely crippled, while the trailer’s cube is very powerful.

If the leader starts off with a good position, he doesn’t double but instead plays on for an undoubled gammon that will win the match. He doesn’t actually double until he reaches some sort of position with very low gammon chances and winning chances a little over 80%. (Typically these are racing and holding games.)

You, on the other hand, can double whenever you have an edge with some gammon chances. You’re hoping to win the match on a gammon.

The upshot is that you can drop and have 32% to 34% chances, or take and have chances in the mid-20% range. (My XG rollout actually put this position at 27%.) So dropping is the better choice.

Huuuugeee thanks for this complete analysis of the situation.
I'll have to think more about money game first as you stated (something I was not doing until now). Very helpful thought process! I started going through 501 essential backgammon problems few weeks ago and really dig it!

Thanks for this amazing book, and for the "free coaching" on this forum