It's a judgement call. I personally feel there is no need to go significantly above 100 flops but some smart people disagree with me. It's really not something easy to quantify.

Buying PioSolver Pro gives 2 licenses. Is it possible to use other licence on an other computer which is not on my name?

I have bought optimal preflop ranges from Fulldesksolutions.com.They didn't mentioned that charts(trees) can be used with PIO(i have CREV-GTO+,that's why i need those preflop ranges because GTO+ is post flop solver).So I tried with PIO free version but when I try to open those files (.cfr) it happens this:
Is this because free version is for turn and river calculation only?
Thanks.

You get one license which can be used on 2 computers. You can share it with someone but from our point of view there is only one license owner. This means the other person has no rights to the license/support/asking for resets and have no recourse if the owner of the license decides to take it for themselves. This means it makes sense to share only with very close friends/family. Scams like selling half of the license and then requesting a reset and taking it for yourself do happen frequently and we will not do anything about it (as we don't have resources to police private deals our ToS forbids in the first place).

I am very sorry but we are not associated with them in any way. To be honest it's the first time I hear about them.

Yes, the free version is for evaluation purposes (it solves turns/rivers as well as one flop: Qs Jh 2h). It doesn't load saves from commercial versions.

Ok,thanks.I knew that could be a problem.They r some guys from runitonce.com but very unprofessional from them not to mention in which form they will supply those charts.Thanks anyway.

It's a good idea to:

1)point it out to them
2)ask for a refund

It's likely they mean well and just haven't communicated things clearly enough.

I have contacted them since i received those .cfr files.I have asked how to open them.They have responded to use PIO.Now I have wrote another email.I didn't ask for refund (55$ is not that much money).Maybe I will invest in PIO in near future so...thanks anyway.

Hey. I was reading about mixed strategies on Janda's book and he said "A miexed strategy occurs when the EV of two lines is exactly equal. In theory, this is an extermely common occurrence in poker because if a player always takes the same line with the same hand in certain situations he can be exploited". So I was in piosolver and trying to be familiar with this concept and appeared a doubt.

Why do on river situations, piosolver doesn't take just one line when in certains situations when involves fold/raise or fold/call and the EV of one line compared to another is slightly different ? I took some pictures to show it better.

https://gyazo.com/928fcc6f01fc3b1b7a2d90cad2d208ff

https://gyazo.com/8f6f9c79b15032d9729e9de7ecfe77f7

with 6h6d the EV of calling is -3.74 and even so the software use a mixed strategy between fold and calling.

with Kh8h the EV of calling is 1.95 and even so the software use a mixed
strategy between fold and calling.

I think this occur only on river, could you explain me why ???

Thanks

The both lines when taken in exact % produces that EV.So 66 if taken only to fold produces even more negative value.

Yes. It's easy to see why: if EV of one action was higher then the other and there was mixing between those those the player could improve by only choosing higher EV action. The definition of the equilibirum is that neither player can improve on their current strategy.

It's because it's difficult to get an exact equilibrium. The EVs are calculated for current equilibirum approximation (which isn't perfect) and therefore the EV of actions may not be exactly equal even if there is mixing. The better the solution becomes the smaller those differences are (well, as long as it's a line played with significant frequency).

Notice that in equilibrium EV of MES (max exploit) is the same as EV of a current strategy. The solver shows what the difference between those. That measure is called exploitability (how much a perfect adversary could win if they adjusted perfectly with opponent staying where they are). The close the number is to 0 the better the solution is.

As to why Pio can't just set the strategy to max exploit (always choose the action with highest EV at given point) the reason is that itself would be very exploitable (once the opponent adjusts).

It's all a bit complicated. It's important to realize the EVs displayed are not some objective set in stone numbers (as those can only be known once we calculate perfect equilibrium). They are just EVs against opponent's strategy in a current solution.

Hey Punter!

The situation I am exploring is HU SRP IP where the IP player has the option to c-bet flop for various sizes. I am wondering 1) Why/how solver chooses a preferred c-bet size option? and 2) If this is the same as me trying to find out which flop c-bet sizes are hardest for our villain to meet their MDF? Does solver take any of that into account or is it best I figure this out on my own? I am guessing that solver doesn't do this but I'd like to know for sure how it goes about preferring certain sizes over the others?

I ask because I've plugged in different sizes and the EV only ever changes by about half a percent which is basically telling me it might be a waste of time to try and implement various sizings on the flop in real time, complicating my strategy for a miniscule gain in EV.

Thanks!

It looks at what the highest EV action is at given point and adjust strategies in that direction. It does that at every iteration. If one size yields higher EV consistently it will over time be chosen more and more often (and maybe always).

I have no idea what MDF is

Yes, adding bet sizes won't increase your EV by much as long as you have at least one reasonable one.

It would more or less a waste of time if you are playing against perfect or nearly perfect opponent. In real life maybe it makes sense to investigate more bet sizes because people are more likely to make mistakes (that is they may be well prepared to play against typical sizes but not at all vs untypical ones).
That is to say you're not only complicating the strategy for yourself but also for them.

That helped a lot, this is what I was thinking as well. Thanks.

MDF = minimum defense frequency. The portion of our range we need to defend to remain unexploitable by villain betting ATC (any two cards).

This concept has nothing to do with optimal play. There are cases where it's the best to let opponent profit from betting any two cards. That happens quite frequently postflop where one of the players has a stronger range. As a simple example this simple 3bet pot spot:
https://pastebin.com/WyNFEhc7

has IP folding to a 50% cbet almost 38% of the time which means betting and (and then auto-folding if called) with any two cards would be profitable. There are many other spots where the effect is even more pronounced.

Another small bugfix is now available via the updater. Issues fixed affected a very small % of users, those are:

1)On some rare trees the solvers produced NaNs (and therefore collapsed) after very long calculations, this is now fixed.

2)Pure preflop trees saves are now fixed (they didn't work but it was unnoticed for a long time because recalculating them from scratch takes 1-2 seconds)

It should be 1.10.8 version for both the solver and the viewer now. Thanks all for reporting the issues.

Hi

1. I want to see only the hands where i have "nothing", and look how many of them I bet. I can go to "range explorer" and check FDs for instance, but it won't show me how many of them I bet or check, as we can see in the main window of the browser. But in this main window, I can't just see only the FDs but the whole range together.

2. How many sizes per Street would you máximum reccommend to use and which ones?
(no rake)

3. Can I round a tree, but keeping the same (aprox) percentages of folds and raises and calls? Because it changes the percentages a lot when you round it.

4. if you save the tree as an small tree save instead of the complete one, should I calculate the whole tree again in order to use node locking and round functions? Because it tells "tree is incomplete"

Thanks

I recently tried installing PioSolver from the original Dropbox link after purchase and when extracting I received the following threat warning. Is this to be expected? Has anyone else seen this?

Trojan:Win32/Peals.F!cl

This is definitely not expected.
It looks like Windows Defender has problems with false positives after recent update. See for example this reddit thread:

https://www.reddit.com/r/sysadmin/co...es_in_windows/

or here:

https://devtalk.nvidia.com/default/t...32-peals-f-cl/

Hopefully they fix that soon. For now you can whitelist the .exe if you trust we haven't uploaded a trojan in our updater or wait for things to clarify (it seems it was introduced in the very recent update of Windows Defender).

Replying out of order because I've missed the post #3491

If you are using 1.10 version you can do the following:

1)Open Range Explorer
2)Select "nothing" and "no_draw" as well as "all" in the top right

It will now show you hands which are nothing and no_draw. It also shows how many combos are selected and what is combined strategy for those hands. For example here:

https://gyazo.com/0845e4e3d0c3fd31b7ace9c319ec0635

There is 15.9 combos of nothing and no_draw out of which you check 15.9 and bet 0.

Trying a cbet in the same tree:
https://gyazo.com/97b5a3b6117cfb233adb84c8475c8a51

There are 19.3 combos of nothing and no_draw out of which 15.9combos bet and 3.4 combos check.

It's a hard question I don't have answer to. Some smart people like a lot of sizes. Some smart people like to make the tree minimalistic to get results faster. It's very hard to answer in general.

Right now it rounds per combo basis so if you round heavily (say to 0-50%-100%) the percentages may change a lot. It's unlikely to change as doing it in a way to preserve percentages isn't straightforward (once you round in one place the range everywhere else further down the tree is affected.

Yes. You can rebuild the tree from a small save (this require executing solver command to do so) but it's recommended to just build the tree from scratch. It's faster and you get a memory check in case the tree is very big.

I'm trying to solve some spot but it doesn't work.

https://pastebin.com/C2d3xUqj

(I want to force IP bet on the flop/turn. Some other spots with these settings work well, but I can't understand why)

I would appreciate it if you could take a look.

It works here. Are you on 1.10.8 version?
There is an option to donk OOP on the turn though which you don't want (if IP is always betting OOP will be always checking anyway).
To remove donks just just leave bet on the flop and donk on the turn empty for OOP, like here:

https://gyazo.com/bc2c3e3fe3033a2e9163deb851d9e0fe

Yep.
When I paste the config, build tree and go I see this https://gyazo.com/ff94dcc984753b1ca4cef0ce2a3f11fa
And this after stopping calculation https://gyazo.com/f18ca4a7e1095dab26debef5e7b2cfc1

Just played with the config. It looks like the problem is in bet sizing area on the turn. If I change IP sizing OTT to 50 (or just add it), it works fine. But if I leave it as 75 or empty and add allin, nothing works.

I want OOP to be able to donk. I want to lock IP always bet facing check on the turn.
So I remove a line Check, Bet, Call, Check, Check
The tree config is exactly what I want.

Yes, I misunderstood you. I thought the problem is in tree-building.
It's in fact a bug. I missed one particular case when removing actions from the tree. It will be fixed ASAP (most likely tomorrow morning).
It affects only very specific trees and it's obvious when it happens (it just doesn't solve them).

What I am saying is that OOP will always check anyway (or just bet with hand which are indifferent but checking 100% wouldn't hurt) if IP is forced to bet all-in.
If you know your opponent has to go all-in if you check then it's optimal to always check. It's true about any bet as long as you have enough raising options available.

Hi,
I am using the version 1.9.2. I "told" pio to update to the latest version but it says it is up to date.
Is that a bug?


One question about script(s)/script generation

Is it possible to create reports when I use the script? Atm I have some space problems on my disk. So I just want to save the trees as "small saves". The problem is afterwards it needs loads of time to create reports for the river.
Is it possible to "tell" the script that it has to create some reports before saving? If yes, how can I do it?

Maybe in the bottom left field of the script generation?!
"Commands to execute on each tree before saving"


If not what kind of usefull things I can make there?

thx in advance

No, we've changed the distribution system to make it simpler. Unfortunately you need a new updater/installer:

https://www.piosolver.com/blogs/news/piosolver-1-10

Hopefully that's the last change of this kind as it can now update itself as well.

It's not possible at this point.

You could save some data but to turn into reports require quite a bit of programming.

There is not much you can do about it sadly. Reports on the river require all rivers to be there somehow. It's possible to implement on our side but it won't be a priority for quite a long time I am afraid. It would also only be possible for reports you specify before running the script as once the data is forgotten you need to recalc it and in case of small saves recalculating all rivers (needed for aggregation reports) takes a very long time (although some of our users still run it).