An interesting problem with many choices I thought.
I'd be interested in hearing your thoughts too Bill maybe once everyone else has chimed in (hopefully)!

Taking he opponent's bar point looks mighty appetising. As does slotting our five point and getting the prep for our board being built for the holding game.

Saying that, anytime you can make the five point in one move you are nearly always correct in doing so, so that could well be the best thing to do.

When in doubt, I think I make the golden point.

10/5, 8/5

i love having the opponents bar point, so 21-18, 10-5. plenty of rolls to make our 5 after this

I make my five. White is ahead in the race. The downside of making his bar is the rolls that hit and cover the blot in whites board. [At the very least, several doubles and 4-2].
Although you don't want to make a strategy of burying your men, there are also scenarios where white can play for the race and bury men behind the blot on the 21 point.
Keeping the 21 point keeps pressure on White, and black will typically roll some hits later in this game.
10/5, 8/5 Easy

Yeah I've just played around with this position a bit and I'm certain making the five is correct. The blot on the bar nearly operates as an anchor in itself, as the opponent has to be so careful about hitting and leaving all the counter shots from both our anchor and their open board with our dangerous looking board (only rolls we really hate are 6/1, 1/1 and 6/6). Obviously this carries a big bonus, as it makes our opponent's life even harder when trying to get their men round, something which is lost if we shift to the bar and let them start burying checkers/hitting our blot easily.

At least slotting the seems a must (if not making it).
Pip count : Black 159, White 133.

We're way behind, so we must maintain contact.
Leaving the blot on White's barpoint seems dangerous, but it isn't that much because White has no board and he must also secure his own blot.

So I hesitate between (in no particular order):

a) 10/5 8/5 --> making our 5-pt
b) 21/18 10/5 --> advancing the anchor and slotting our 5-pt
c) 18/13 8/5 --> securing the blot and slotting our 5-pt

I'm not sure about advancing the anchor since it allows White to play behind it and attacking the back man.
If we keep the actual anchor, no need to let the blot on the barpoint and risk a disastrous 1-1, so I would move it.

18/13 8/5

P.S. I assume it's a cash game (with the usal Jacoby) context.

Btw, the golden point is our opponent's 5-pt, not ours.

Not a bad aurgument. Maybe I'm too scared about game-changing rolls like 1-1, 6-6 and 6-1. Simply making our 5-pt might be best afterall. White could leave a shot as soon as next roll so having a 3-pt board might be handy.

Over the board, my thinking would go like this:

Three plays -- 21/18 10/5, 18/13 8/5 and making the 5-point.

I don't like leaving the blot sitting on the bar, so the 5-point is out and we're down to two choices.

The bar-point is usually a slightly better holding point than the 4, so 21/18 10/5 is appealing.

If my remaining blot weren't on the 21-point, I'd make the bar in a shot. But that's an awkward spot for a blot, and White can attack it. So maybe look at the position after 18/13 8/5 a little more.

Better count the race -- I'm down 18 after the play. Nothing there -- either holding game plays well at -18.

Aesthetically, I know 18/13 8/5 looks prettier. But I've had a lot of practice games against Snowie over the years where not making the bar-point in similar positions gets flagged as a modest error, no matter the aesthetics.

OK, so I play 21/18 10/5. Oops. It's a doubles match and my partner wants to make the 5-point. I punch him in the nose and pick up my dice. Tough game, backgammon.

I've set up a rollout to run while I go to the gym. I'll post the results mid-afternoon. If someone wants to do an XG rollout and post it at the same time, that would be instructive.

I'm running an XG rollout will post when done.

IB

Sigh I'm so dumb. Oh well learn something new everyday. Can look slightly less ridiculous in front of good players than I would have yesterday now

I don't see a big benefit to making an anchor on the bar point. You already have the anchor on the 21 and would have to break it to make the bar point. I just play 18/10.

Oops, yeah, 18/13, 8/5 is clearly better than 18/10.

I don't understand the 13/10 part of your move. The 10-pt seems pretty useless at that point, no? Why not slot the 5-pt?

As a side discussion, if we make the bar (21/18 10/5), does White have a double?

Before I read through the replies I thought this through and came up with making the 5 point. My reasoning was similar to Wamy's above: white's got a wide open board, and black can close his down to 3 points. The only awful white rolls for us are 66, 11, 61, and 22 (and 22 is not even THAT great for white). Any other kind of hit will have to leave a direct shot somewhere. Meanwhile, we've secured our own 5 and have a good looking home board.

After reading through the replies, I can say that all of the other options you all brought up are convincing as well. It will be interesting to see the rollout results.

This too is a good question. I'd say no, because we still have a lot of volitility left in the game, the position isn't completely dominating, and the lead isn't that big. However, proper doubling is currently my weak point, so I'd like to hear answers to this question as well.

I'd say no as well. The only thing white has going for him at this point is the race. His home board is weak, and he's even got a blot on the 11. When black makes the 18, I think this keeps white from doubling for quite a while. I think white won't have a double unless he starts making some points in his home board while also hitting black's extra blot back there. His chances of priming black's lone checker in are gone now that black took the 18 point. Black could too easily escape that back checker, hold on to the 18 to defend, and be building up a strong home board himself. So I say white is really in no position to be doubling right now.

First rule of checker play problem solving...
make sure you consider ALL the candidates.

Slightly surprised by BR's last post. Don't know what the rollouts will say, but once you make White's bar point, White's strategy becomes far too easy: Attack the blot on the 21 point, and play for the race.

Black making his 5 and keeping the 21 point, keeps White from attacking to a large extent, and sets up a better chance to hit later.

Well.. no one even contemplated the best play according to this strong XG 4-Ply rollout:

I'm really astonished that 21/18 10/5 loses almost twice as many gammons as 18/15 10/5.

Btw, does that XG rollout take into account the Jacoby rule? And how could you induce the "Jacoby" parameter into a rollout in the first place? I mean, it must not be trivial. Right now gammons don't count (assuming Jacoby), but they will once the cube is turned. I guess that the rollout applies the theorically double/take/drop actions along the way...

and that's why bots win, no human would ever make this play here

Yeah I'm struggling to even see how you could frame this move in a set of logical rules that a human could successfully apply to the game.

Moving 18/15 isn't a running play, running is futile, it's a confrontational play forcing a hit on some 4's (and a bonus with 22) with a view to re-entering the checker as a rearguard with the subsequent nuisance value.

rearguard, I love it! It's basically playing backgammon like curling! Like higonfive with his football analogies, I'm gonna incorporate curling terminology in my backgammon.

Forget about checkers, it's now stones!
Owning the cube? It's now "having the hammer"!
Slotting? What about drawing?!!
And a hit is... still a hit!

You want more?

Our home board? Now it's our house!
Pick and pass? Now it's hit and roll!
Pointing? What about freezing?!!
The box in a chouette? It's the skip!

All right, all right, that's enough.