Black's up by 16 pips. Is this a good double by black? Is this one of those "action doubles" maybe?

Hmm, I believe white has an easy take given that black still has 3 men back and he should be able to secure an anchor soon. It looks like black would need a very favorable sequence in order to lose his market (an average sequence would be something like black makes his 5 point and white anchors, after which white has an obvious take), so my guess is no double / take.

Yeah Gnu rollouts have no double, take (double/take = -0.106). As dsaxton says, White is likely to anchor, after which the position will be close to even.

Don't post rollouts so soon, please.

Anyway, I think this is a nodouble.

oh well, my first idea watching teh table was almost it was toogoodto ... if doubled in white i would pass anyway . too many gammons i think. but if gnu said that ...

Looks like a big no double and an easy take. In the most likely sequence, Black makes his 5-point and White enters. Now Black has to get three men around the board while hoping nothing goes too wrong in the process. That's going to take a lot of work, and plenty can go wrong in the process.

If this isn't clear, the best thing to do is to set up the board and play out a few sample games. Notice how much work Black still has to do, and how often things go wrong.

One downside of the bots is that players lost the habit of playing positions manually to see how they develop. Bots can give you the right answer (as here) but only manually playing the position gives you the experience to "see" why the answer is right.

To be fair, if you look at the bot's answer here, I think you can quickly figure out why that answer is correct, without needing to play it out. I get what Robertie is saying though, hands on can be very useful.

ok, i guess that's the right scenario. but i see also a lot of chances of gammon here. black can easily hit the second cheker , or white could easily hit the board again.
I don't want to refuse gnu results : obviously this is a bug in my thoght process, beeing too much aggro with cube.
Just a question: do the results changes if black ownes the cube ?

Haven't checked, but cubeownership should make black less willing to double. Owning the cube is valuable to you since it prevents white from doubling if your position should take a turn for the worse, so you usually need to be a slightly bigger favorite to redouble than to make an initial double.