I tried to do research by asking more experienced people in the niche. I've got many conflicting suggestions and advice so I went with my gut feeling. I felt there was some reluctance in the gambling world against subscription model and I felt one time fee would maximize chances to get it going.
As to the cloud vs stand-alone dilemma, the reasons for choosing stand-alone were:
-everyone has a CPU at home and those could be utilized; building CPU heavy infrastructure so people could run it there would be very expensive and I have 0 clue how to go about it; it seems there are no CPU-heavy reasonably priced options on AWS for example (most instances are geared towards network/storage/RAM but there are none which are both decent when it comes to CPU performance and RAM, the ones which are the closest are very expensive)
-people who play for significant money don't trust cloud in general; especially for high stakes players it's important that no one could access their calculations
As to one time fee (and then maybe another for 2.0 version) vs subscription for updates, I feel maybe subscription for updates would be a better model but it was easier to start with something simpler. I had no idea if it's going to take off and I didn't like the idea of very few buyers demanding updates after 6 months or so when it's clear the product flopped (it didn't happen but it could).
It was a difficult decision though. I still don't know if I've got it right but I feel I did (some competition who went with cloud model didn't do as well from what I understand).