It depends what the turn scenario is (how many bet sizes, what stacks etc.)
Better way to go about it is to give me your CPU/RAM setup and then I can give you some approximations. Post here or PM me for details.

Lets say 1 bet size, 50bb (same parameters of the turn). I have a i7 860, 6 Gb RAM. Its ok 6 Gb for a flop with that parameters? What can i do with 6 Gb?

6GB is usually fine for 1 bet size @ 100bb flop+

With 6GB you will have problems making more than 4.5-5GB available. That will be enough for 50bb trees without multiple bet sizes, even if they are small. A tree with 50% cbet, 65%/75% turn/river bets and donkbet allowed is 4GB. That's probably the biggest one you can comfortably solve. Things like 100bb spot without a flop donkbet or 3bet pot will be way more comfortable.
As to 100bb trees: a lot depends on bet sizing here (the smaller the raises the more branches for example). You can usually fit 100bb tree with all options allowed if the bets are reasonably big.

"Include oop bet if called previous street" << but if previous street was check-check and I want to consider OOP bet but not all in, then what can I do?

All other OOP bets are included by default. The only ones which can be disabled in standard GUI form are:
-the ones after check-calling previous street
-the very first one in the game

If you don't want to consider all-in just uncheck "add all-in" checkbox on this street for this player. Maybe the best would be if you paste a screenshot (or pm me) and tell me what you want to achieve.
If you want to make some specific tree with some untypical branches disabled you may need to drop down to scripts for now but I think almost all real life cases should be possible to insert from GUI.

okay here is gui, oop range is here:

AA,Qd9d,Qd8d,Qd7d,Qd6d,Jd9d,Jd8d,Jd7d,Jd6d,Td9d,Td 8d,Td7d,Td6d,9d8d,9d7d,9d6d,8d7d,8d6d,7d6d

IP range is KK

the problem is when the turn goes check-check and the river is for example 3h, OOP still checks 100% and IP also checks 100%.

This is a bug in our tree-building form which is now fixed in dev and will be updated in the next version.
The way it is now is that at least one line on the river needs to end with all-in.
So you can add a raise size on the river (say 100%) or reduce effective stacks to 100 (but then no 2nd barrel).
I think the closest you can get to it for now is to add 100% raise for both player on the river. It's not exactly what you want probably (because there will be check-check/3h/check-bet-raise available)

I am sorry about it, we didn't predict that trees without raises are sometimes interesting to look at and we didn't test for those cases.

Just ordered software yesterday and am super impressed so far. One thing I wouldn't mind seeing is the ability to use just a few brackets for mixed strategies. For example, instead of having a hand be 94% bet/6% check, forcing the program to use say either 0%, 25%, 50%, 75%, or 100% (or other arbitrary numbers like 0/33/66/100). Seems like this would definitely make strategies much easier to implement for a human, which I think would certainly offset the deviation from actual GTO strategy.

Though I do see how trying to bracket hands could certainly throw overall frequencies way off. Just trying to brainstorm for ways to make the results more easily implemented in a practical sense.

This is one of the most common requests we get. It requires some work though as classical algorithms are based on the assumption that you can adjust the strategies very slightly every iteration. There are algorithms which are not based on this assumption and we will certainly try some of them once we are back to optimization/algorithmic work.
It's not likely to make the top of to-do anytime soon though as there are many very nice features to add before we attempt this one.

We are adding a lot of features to make solutions easier to understand/visualize but the one you suggested is not among them for now

Hello,

Maybe I don't have strong understanding of GTO as you and some others, can you explain to me why the strategy that it comes up with is GTO? My problem is that we are using a fixed starting range (which hasnt been proven to be GTO), a fixed betting strategy (which also hasnt been proven), and parameters we use to simplify/limit the game tree. Or is this software calculating the best strategy for the situation we create (range, bet size, etc..)? Thanks

We are solving for equilibrium with those assumptions. This is a post-flop solver (for now) so it solves postflop games with the assumptions you mentioned (ranges from preflop and some betting abstraction).

As to the bet sizes: yes this isn't full NL game with all bet sizes possible at every point. It is however very close as simulations show that adding more doesn't change overall EV much.

Yes, it solves for equilibrium in a bet sizing abstraction you create.

How large are the solution files, in PioSolver?

I really like the interface better than the competitors, and it's coming down to solution size, in choosing a product:

I was talking with the creator of a pseudo-competitor, and they claim(ed) that PioSolver are quite large, in the gigabytes, while the other programs have smaller solution sizes/require the turn and river to be re-calculated every time, and so on.

If the solution sizes are, in fact, still in the gigabytes, then would PioSolver be willing to add this option for a smaller solution size, with re-calculation, if I bought the Pro Version and then suggested it, and so on?

Also, Simple Post-Flop has this nice feature, where they let the user select a blanket gate like "hearts", and so on, when constructing a range. Would PioSolver be able to add a similar feature to the next development issue, if I bought the Pro Version and then showed them what I was looking for?

For now we save the whole tree (so huge) but in the next release (hopefully less than 2 weeks from now) we will have small saves. You will be able to choose either flop+turn save or only flop save.
Flop + turn is a recommended option which will produce files 50x-150x smaller than the full ones and barely noticeable (usually around 200-300ms) delay on the river when you browse (it calculates those on the fly if you made a small save).
Flop only solutions will fit on the floppy disk although they are not a good idea in general (cause recalculating turns is painful).

They are correct for now.
Again, recalculating turns every time you enter it isn't the most usable way. The best way to do it in my view is this:
-if you solve the tree you have access to all of it and don't need to recalc anything
-if you save the tree you have a choice of saves: flops only, flops+turns, full
-if you load an incomplete tree it recalculate turns or rivers (depending which one you've chosen when making the save) on the fly and then it remembers the results (so you don't need to recalc again when browsing)

This is exactly how we implemented it. Flops+turns saves are the best practical solution by wide margin because they don't ruin the experience when browsing the tree and are still between 50x and 150x smaller than full trees. Also you don't need to recalc anything as long as the whole tree is in memory.

Yeah, it's working in dev, like this: http://i.imgur.com/doMrcBj.png

We did overshoot a bit with new features in the next release that's why it's delayed a bit. Still should be out in 2 weeks or less if everything goes well. Here is very incomplete list of things in there:

-small saves (choice of flops, flops+turns, rivers)
-reports exportable to .csv format (you can list all the board runouts for the line you are interested in and it will show you action freqs/evs/eqs for every card, so for example you will be able to see on which turn card 2nd barrel is made the most often or on whcih turn/river card 3rd barrel is made most often etc.)
(here is an example of such report for 3bet pot sorted by 3rd barrel frequency: https://cdn.shopify.com/s/files/1/07...05130986031534)
-setting/locking strategies in chosen nodes
-range analysis (so breakdown to stuff like top pair/underpair/set etc. for any point in the tree)
-equity calculations (for both the tree or arbitrary ranges)
-instant evs/other results even for humongous trees due to result caching

All of this is already working in dev and we are just polishing/testing those before the release.
There is much more but those are major points.

Can we compare equities of any two branches? For exampl equities of turn cbet range vs folding range of villain?

Yes but in the first iteration you will have to copy-paste yourself which won't be a problem as you can open several instances of the range explorer and use them along when browsing the tree.

Actually this is one of the cases which can be done automatically without copy pasting (it will be possible to open the range explorer in any node, so just choose fold and open it there). It will also be possible to make range explorer "follow" the tree browsing and show results/range distribution for every node you are in.

when are you going to release next update? I want this feature very much cause in many cases I am interested in river situations when turn goes check - check.
p.s its awesome soft thanks a lot.

It's very likely Tuesday or Wednesday this week.

You will be able to force check-check on the turn now by just removing bets/raises from the turn although I am not sure if that's what you really want

I didn't get it well, so what I really want is that after I make turn goes check-check like in this picture
http://i.imgur.com/6AaFMN1.png
software doesn't consider beting the turn and is checking 100%.
Both of the check box "Include oop bet if called previous street " are selected

http://i.imgur.com/uiGQSSW.png

p.s I tried now and when I select the check box "Add allin" now it works but what if I don't want an all in option?

I think you meant "doesn't consider betting the river" and yes, this is a bug I mentioned. It works in new version.

Yeah, if you add raises it works but in the new version it will just work with any bet sizing options you choose.

Hey punter I was wondering if you could explain something to me that I'm just not getting. In some river spots(maybe this occurs elsewhere but this is usually where I see it) the solver will c/r bluff with hands that are -ev.

As long as the solution isn't perfect equilibrium there will be some plays which are -ev (or not the highest ev). There are smaller and smaller mistakes as the solver progresses.
That being said it's possible that for hands which are in given branch very rarely those mistakes are still sizeable comparing to overall exploitability.
Feel free to paste some example, maybe a screenshot, preferably with overall exploitability of the solution so I can take a look if it's something unexpected.