Quote:
Originally Posted by Powers_That_Be
Can you compare/contrast this to StoxEV?
Yes, although I have to disclaim that the following is just my personal opinion. I'll do my best to be fair.
Both products are designed to model poker situations. Both products take as input poker situations, (hand ranges, pot sizes, bet sizes, etc) and output EV's and strategies.
Here are the advantages of Stox EV:
1. Models a wide range of poker situations
2. Can determine the way to maximally exploit an opponent model. (This is sometimes called a Bayes response)
3. Can determine Nash push-fold ranges over 1 round of gameplay.
4) Manual exploration; you can input two complete strategies and play with them to see how adjusting strategies changes EV's, etc.
Here are some disadvantages of Stox EV:
1. Depending on complexity of model, often a long input time.
2. Bayes responses can often be very risky, for an extremely simplistic example, if the opponent folds 51%, your tree might tell you to bet pot with everything. If he folds 49%, you might only value-bet. Both strategies are also highly exploitable, so if you are even a little wrong about your read, you could wind up losing a lot.
3. Can't really find GTO strategies for most situations, although if you are diligent enough you can find something close via manual exploration.
Here are some advantages of Hold Em Forge:
1) Able to find Nash-equilibrium calling ranges within the modeled situation over up to two rounds of action. (Unexploitable, safe, but not as profitable)
2) Able to find Bayes response calling ranges within the model over up to 2 rounds of action. (Maximum exploitable, highest profit potential, highest risk)
3) Able to find restricted responses within the model over up to 2 rounds---these are like a mixture of playing unexploitably and exploitably that keeps your risk (the maximum exploitability of your strategy) below a user specifiable threshhold.
4) Input often as simple as just loading two hand ranges from file and pressing go.
5) Novel output displays display EV's and strategies super intuitively on your screen---you can see them in a pokerstove-style output box color coded by strategy, or graphs of your play by hand type. There's also a handy fast facts box that can makes critical information like your worst calling hand easy to find.
and the disadvantages of HMFA:
1) Software only designed to model continue/fold decisions (and the push/check decision on the river); the more options you allow both players, the less applicable HMFA's analysis becomes.
2) You can't see the entire tree at once; only one street at a time. You have to use the feed forward option to see what to do on the second round of the analysis.
3) Software designed for heads-up play only.
So the bottom line is in my opinion the following. HMFA is great for understanding baseline, GTO ranges, and seeing how these ranges change as you put more or less trust in a read of the opponent.
Stox EV is best for modeling complex situations in which you are considering a lot of action sequences, like a tree in which you bet some hands, get reraised sometimes, called sometimes, folded to sometimes, etc.
I would say if you are just learning poker, you would probably prefer HMFA over Stox. This is because you can build an intuitive understanding of what the sort of default, GTO plays are in a ton of spots in just a short time with HMFA. The fact that HMFA comes preloaded with common preflop NLHE ranges also helps. If you are an advanced player, I think you would spend more time using Stox, although you would probably want to get both products. That's because HMFA can give you a quick and easy answer to some questions Stox can't, and vice versa. (IE, HMFA could tell you what's a balanced range to not fold against a third pot bet while Stox could tell you the EV of raising say two pair+ if the opponent will shove over with x range and fold with y range).
I hope that helped, and again I don't want this thread to turn into a war between stox and hmfa; I think in the end they are each good at different aspects of poker problems.