Can I run GTO+ with the following hardware:

- Laptop
- i5 3230M 2 x 2.6GHz
- 4GB RAM (3.89 GB free)
- 64 bit system
- Windows 8

Would like to purchase the software, but need to make sure that I can use it on my laptop. Thanks

The number of cores, processor speed and memory are a bit on the low side, but there's still a decent chance that the software will run at an acceptable speed. You can test this for yourself with the trial version. For example, you can try the following testfile: www.crevfiles.com/gto/savefiles/testfile.zip. On my own system, which is a bit above average, this will takes about 35 seconds to solve to 0.5%. On your system I expect that it will take about 2 or 3 minutes. Should it be significantly longer, then your laptop is probably an older model. The problem with systems that are older than ~5 years or so is that at the time processors were not yet as capable at quickly performing operations on large chunks of memory as they are today. GTO software requires some of the more recent advances in hardware (about 5 years ago) in order to run smoothly. It's not possible to tell this from the specifications, so please just test the solving speed for yourself.

Thanks for the quick response. I took me 240 sec to reach 0.5%. So probably my laptop is too slow to run GTO+?

It does indeed seem to be a bit on the slow side, although it's not the worst I've seen.
A more recent system with better specs should show a far better performance.

Thanks Scylla for your patient answer.

Im making this comment in hopes to add some things to GTO+. +1 if you agree.

Because of the awesome database mode I am solving batches of trees with different board textures while I'm at work and then I can come home and evaluate strategies for those textures when I come home. Also when I review my sessions I will search my database for trees/textures that mimic each hand instead of solving each individual hand i want to review. But there is no way to organize these trees/textures and sifting through 500 trees to find the board texture most similar to the one I'm looking for is tedious and time consuming. I would love to be able to either search for certain textures within my database in much the same way that i can specify a board texture in CREV. Either that or I would love to be able to organize them myself (cut and paste, etc.).

Secondly I would love to be able to export a certain tree in my database to a new file in order to node lock and solve for exploitative adjustments. Either that or a function that allows me to somehow use the node lock function within my database in a way that would allow me to compare the new solution to the original GTO solution.

I think the race to become the most effective/efficient solver will be determined by the one that makes studying quick and practical. Thank you Scylla for all you've done. This product is the nuts.

Scylla，Does it possible use GPU to solve the game tree? I think it's a competitive feature because same price GPU with CPU usually has 30+ times performance on calculate.

If not possible, will add thread numbers on future? my CPU have a 10 core 20 thread.

(As a bulletin for other GTO+ user：I use scylla's test file solver to 0.5% use 20s. software only can use 16 thread)

www.crevfiles.com/gto/savefiles/testfile.zip

+1

I actually logged into My account here in order to suggest the same thing. It would be great to be able to filter the database by board texture (Axx/two broadway/no broadway etc) Currently im having to do it manually myself using pen and paper to see which board textures are good or bad for my range.

2- Also im not sure if i am missing something, but is it possible for the node locking feature, for us to have suit selection like we do with the preflop ranges here?

http://prntscr.com/j5ikxh

it would be so much easier for me to node lock checking/betting BDFD etc a lot quicker if i can just quickly select the suit then click on the hands i want to apply it to.


- The software is great though overall. Other GTO software can no longer justify the huge prices they charge.

Regards,

Smasher.

Ok, we can look into adding some sort of filtering options for later releases.


Ok, I'll see if something can be done here.

Currently the max number of threads has been set at 16. This is however only to protect inexperienced users from entering a number such as 1000 and running into all sorts of issues. You can set a number higher than 16 by using the code XXoverride. Here XX is the number that you're looking for. So if you enter 20override your CPU usage should go to 100%.

For this right-click "2crd bckdr fd." to fix the relevant hands in the table and matrix.

Hello Scylla,

I have recently bought a license for your software CardRunnersEV, very impressive all the features that it offers!

I could not find a response to my question below on the forum or in the tuto, so I take the opportunity to contact you directly:

I have imported the hand below (Hero is MP+1 and Villain is Cutoff) and modified it to see what I should do when facing such a raise on the flop (I folded in my original hand).



On the flop, the software shows that faced with a raise, hero 3-bet 100% of the time. Is it the right play or did I make a mistake in setting the hand?

Thanks

you given CO only bet option then CO bet his all range, same as MP+1 a 100 raise range.
then CO response a 100% 3bet range to exploit him.
try add full action option or click right button on position icon, auto build a full game free.

You have only entered a rather limited game tree where, for example, at the start, MP+1 and Cutoff only have the option to bet and then raise. It's better to use a full game tree with all bet/check/raise/fold options. To create such a tree in CREV, right-click MP+1's first decision to bring up the tree wizard. After that, run the GTO solver to let the software figure out the optimal way to play for both players. You will need to fill in a range for MP+1 as opposed to a single hand though. For demonstration videos on using the tree wizard and solver, please go here: http://www.cardrunnersev.com/download.html

That being said, I think you may prefer using GTO+ for this, which is our new GTO solving software. A license for GTO+ is already included in your purchase, so you already own it. GTO+ is specialized in this type of analysis, and has an interface dedicated to this purpose. For more information on GTO+, please go here: www.gtoplus.com

Thanks Jasonhau and Scylla! It works, much appreciated!

What config do you use ?

The particular system that I'm referring to is about 5 years old, and has 4 cores of 3.5 gHz.

Hello, I just purchased CardRunners EV this morning. I'm pretty new to poker but I'm studying hard. I play on America's Card Room (Winning Poker Network) and I have Hold 'em Manager 2. When I copy a hand from Hold 'em Manager 2 and paste it into Cardrunners EV, I get a failed to import hand history error. I assume it's because WPN isn't an accepted version, but is there a way I can convert them? It sure would be easier than entering them by hand.

Thanks in advance for any assistance anyone can provide, it's greatly appreciated.

Sorry, thanks to a really quick reply from their support, I was able to figure out what I was doing wrong. It turns out, I'm an idiot. Sorry I didn't think of the support email before I posted.

Is it possible to study flop bet sizing strategies with Crev without building the whole tree until the river?

As as example:
-BTN Opens
-BB defends
-Flop is A72r
-I want to see the EV when betting 33%, 67% and 100% pot without building the whole tree until the river

I know that the above scenario is possible when we want to see the equity but can I also see EV?

It's indeed possible to enter a tree with multiple bets in CREV. The tree should be entered all the way up to the river though, in order to ensure that all runouts and scenarios are taken into account. Creating such a tree does not require much effort on the part of the user though, given that it can be built with the tree wizard. After creating the tree, run the GTO solver to figure out unexploitable play for the two players. This solve should typically take only a few minutes (depending on the size of the tree and wideness of ranges), after which you will have the Nash-solution for the spot that you're interested in.

You may actually prefer using our new software GTO+ for this though, given that it's specifically aimed at this type of research. With the tree building wizard in GTO+ you can build your tree within minutes and then run the GTO solver to get the Nash equilibrium for that tree. For more on GTO+ please go here: www.gtoplus.com. For demonstration videos on GTO+ go here: http://www.gtoplus.com/videos/. The tree wizard is described in video 3.

GTO+ is included with CREV (and vice versa), so having either product will mean that you also own the other.

Hi Scylla,

I am really enjoying the product. Would it be possible for GTO+ to accept the Pio format for weighted ranges? These are of the form:
AA,KK:0.75,QQ:0.75,JJ:0.5

Thanks

Yes, this should be available in one of the future updates.

I am using the equilibrium solver of CREV to simulate a 3bet IP vs UTG call scenario.

The board is: AJ2r

For this purpose I am using the following ranges:

UTG calling range vs 3bet:

https://gyazo.com/5d98addbe394b033aa638ccbf32e564e

IP 3bet range:

https://gyazo.com/8793575aaabc73cc24c59d1c8e371149

I am solving until 0.5% dEV with the restrictions that UTG can not lead the flop and the IP player bets the flop for a 1/3 PSB. These are the results that I get from CREV:

https://gyazo.com/229ca923cd9ad8df5c14217aec05f447

Following results doesn´t make sense:

- IP betting freq. should be way higher than 60.5% when using a 1/3 PSB OTF on a flop that favours our range
- check raise freq. of UTG should be way higher than 10.6% when IP bets for 1/3 PSB

The same scenario with identical ranges in PIO:

- IP betting freq. OTF close to 90%
- check raise freq. of UTG around 23%

The PIO results make sense to me. Where did I go wrong with CREV?

CREV and pio give the same results; in most cases differences turn out to be due to the trees not being the same (in others due to differences in ranges, switching IP and OOP, etc). I can't tell enough from just the screenshots though, given that they only show part of the tree. Can you perhaps send a savefile to support?