maybe in this case because the counter was [probably] unexpected it was better to attack it last.
Exactly. The ambush is likely to be a surprise because the picks were later (pick 3 in the target bonus and 4 in the ambush spot). Instead, consider if my picks were 1,2 and I only got the ambush spot. Then it's possible the opponent also made those picks 1,2 and they have intel on my ambush. It's still possible they have intel on the ambush or may try to counter it blindly.
Another reason I risked the late attack was because I had a safe ftb elsewhere. If I did not have that, I would have probably attacked with 4 or 5 first order.
I might have used a random number generator and had attacked last turn with a 90% chance. The reason is that with my 1 more income and the extra armies killed, I could just hammer the bonus next turn then and probably win the game by default.
Very powerful decision technique and supporting reasoning. Probably the ambush will succeed first turn most of the time, but we're not in a bad position even if it fails.
Note that the ambush is a from a wastelanded bonus. Let's say the ambush failed and the opponent completed the bonus. Now, if they deploy all next turn and clear out the ambush spot, they have a big stack all dressed up but nowhere to go.
Edited 7/24/2017 20:08:08