PioSOLVER doesn't do fractions in pots/stacks. Just multiply both by 100 and it will work.
This tree is going o be quite a big one with all those sizings btw, it's highly recommended to not use 2 raise sizes everywhere, especially when you are starting.

Alright clear, thank you.
I shall start at least for turns and rivers with lesser options.

Yeah, that's a good idea. While it's really memory efficient now and you will probably be able to build those trees with 2-3 sizes everywhere they still take long time to solve and there are diminishing returns when including thigns like 2 check-raise amount for OOP on the turn

can we please get an option to specify a donk size on the turn? (and maybe only allow it on certain cards eg middle card pairing). currently if you want to allow for a turn lead after x/c you have to allow all turn sizes and this blows up the tree real quick.

Jesolver update: The much-requested smaller small saves are finally there. Some numbers:

Yes, added to the list. This is quite often requested actually.

This one won't happen. We made a decision to force symmetric bet sizes on every turn/river runout. The reason for that is that it makes many features possible (like aggregation reports, line frequencies, hotness etc.). Those features rely on the fact that you lines are symmetric no matter what turn/river card fall. We think that's a good trade-off.

Jesolver is now also 100% compatible with 1.8 version. I recommend you guys check it out as it's really something special when it comes to speed and now storage size as well: http://jeskola.net/jesolver_beta/

Hi, I'm considering upgrading my processor. Can you tell me if piosolver would benefit from xeon processors with like 14cores/28threads? Or does it need more clock speed? How important is clock speed vs number of cores?

Rake is not supported yet though (working on it).

Assuming those are CPUs from the same generation it's more or less number of cores * frequency with small performance penalty for big amount of cores (which is less than 5%).

Hi,I really appreciate some help with node locking.

The spot is T88r BvB SRP.OOP can either bet 1/2p or check.I want to lock that OOP cbets JJ+ at 100% frequency,but I don't want to make any further assumptions either about OOP's cbet range or his checking range.In other words,I want PIO to figure out what optimal play would be considering OOP is forced to cbet JJ+ 100% of the time,is that possible?

Yes it's possible. Combo-locking was introduced in 1.8 version. It's described here:

https://www.youtube.com/watch?v=gJEPGSIpIBM (point 12)

Thanks for the reply.Is there a chance that there might have some bug with it?

What is happening is I lock the betting strategy,set it as "fixed",then go and click on "lock selected combos",paste the same exact range I've put for the betting range.When I click on "set strategy and close" and run the solver,the solver considers as if I hadn't locked anything at all.

Trying to see what caused it,(not sure if that's the problem,just pointing out what I realized)after I set everything up and click on "set strategy and close" and click ctrl+n right after it,the option "fixed" I had selected before for the betting range goes back automatically to "proportional" without me doing anything to cause that.How can I fix that?




Also,I have another question


1)I'm able to make a script if I set the same parameters for all boards I want to solve,but is it possible to make a script with different parameters and ranges for each spot?

There is always a chance there is a bug but it's very unlikely.

Don't do that. Just mark the combos which are supposed to be locked leaving the ones which you want unlocked unmarked. In your example just mark JJ with weight 1 and leave the rest.

This is normal. Those are just for editing ranges. The end result is visible on 13x13 grid in the main window. You may want to watch a video about node-locking, here:

https://www.youtube.com/watch?v=a_OhaxBi7mY

It was made before combo-locking was available but it explains how the interface works.

Yes. Join our Skype group and read the group FAQ there. It's explained how to do it there. You can read the FAQ here.

Thanks for the reply.

Not sure if It's a mistake I did,but combo locking only seems to work when I set frequencies of "1" for all combos I want to node lock,and if I set a mixed strategy for any combo(e.g. 0.9) then the problem I described on my previous post happens(solver ignores locked combos when solving),is that right?

Watch the video with combo-locking example carefully. There are two things to set when combolocking:

1)the strategy (which you do the same way as normal node locking)
2)choosing which combos should be locked (this you do by clicking a small range selector icon at the bottom of the node locking window next to "lock selected combos")

In point 2) you only use 1 or 0 for every hand which means lock or don't lock indicated combos.

It might be confusing at first but those are two pieces of information which are necessary for the solver to know what you mean. Again, try following this:

https://www.youtube.com/watch?v=gJEPGSIpIBM (point 12) carefully.

Hi, I am a new user, likely noob questions...

At the tree, when I fill for example [Starting Pot: 34500], [Effective Stacks: 119728] Pio returns "Server Error: ERROR: set_eff_stack incorrect or missing argument", if I reinsert a smaller effective stacks likes 55123 build tree ok. At some point of larger effective return this error. For calculate this tree I decrease proportionally values and so works fine.

I activate the setting [stop calculation if desired accuracy reached: 50 % of the pot], is it a good % to analyze the spot? I got worried, because the calculation finishes in few seconds, very fast.

About lines OOP we have check / bet sizes, however how as I see at some spots a raise after check (check-raise) is better than bet, do you have some good idea to compare these actions?

Yeah, max stack amount is 65535. I was a bit paranoid about memory usage when making that design decision.

50% is very bad
0.5% is a decent overview, 0.35% should be good for most practical purposes. 0.25% contains very little mistakes.
I don't understand the question. What actions would you like to compare?

That's what I'm trying to point out.This^ means combo-locking does not work when I want to set a weight for a combo that isn't either 1 or 0 (eg. 0.9),right?



Apologies for all these questions

It does work.
Again, there are two pieces of information:

1)what strategy to set
2)which combos to lock

the small range selector near "lock selected combos" is responsible for 2) (and that's why you only use 0 and 1 there).

You use normal node-locking interface for changing the strategy. Try it:

1)solve a very simple tree for a minute
2)stop the solver
3)go to node-locking
4)choose "fixed" for a bet and change JJ to 0.9
5)click a range selector next to "lock selected combos" and choose JJ there, like here:
https://gyazo.com/e760d6d7102171c8d296032df9e3533d

6)click "set strategy and close"
7)verify that JJ is now betting 90% of the time and that you see COMBO_LOCKED in the node description, like this:

https://gyazo.com/64dc7e84e8d6c388f9dab87b3dc85287

8)Click "go" again. The solver will now not change the strategy for JJ

I hope that makes it a bit easier

haha gotcha gotcha,tyvm for the help

I'm glad I asked here,if I hadn't I'd have missed such an awesome feature!

Hi,
Can i broswe the preflop solution (which is sold on PioCloud) w/ the PioSolver Pro?
thanks.

Not at this point as browsing requires recalculating a lot of things and those can only be done by a preflop solver. It may change in the future though.

So i should go upgrading to the edge version first

But i don't need that much ram to broswe it,right?

There's now a Linux version of Jesolver. It's not extensively tested yet but seems to work. You can get the free edition from https://jesolver.com/download/free/linux/

While it's not possible to lock general assumptions like "4bet max 3-4%" it is possible to choose a specific range.

There are two ways to achieve it:

1)Use node-locking feature:
https://www.youtube.com/watch?v=a_OhaxBi7mY

2)Build a tree with a 4bet only and choose a 3%-4% range for the opponent

The second option is going to produce skewed EVs (as all the folds will not be counted) but they will produce the same strategy.