GTO+/CardRunnersEV?

Hi scylla,

is there a way to import ranges from Flopzilla Pro or GTO+ to CardRunnersEV?

You can import ranges from CREV and Flopzilla v1 into FlopzillaPro and GTO+, but not the other way around.

hi scylla,
I'm currently using an old (very old 😃) computer for my sim ( proc : I3-3220 @3.3 Ghz (2 cores 4 threads) with 16Go DDR3@800Mhz. For example with the testfile found in http://www.crevfiles.com/gto/savefiles/t.... I run this sim in 32s.
I'm going to upgrade my computer with an I3-12100 (4 cores 8 threads) or I5-12400 (6 cores, 12 threads) and 16Go DDR4@3.2Ghz. COuld you roughly estimate the gain in time than i can expect ? *4 faster,*8 more and do you there will be a great difference between this 2 candidates (more than 20%)

Of the two CPU's I would go for the I5 one (as opposed to I3), with 6 cores.
I would expect it to be roughly 3 times faster than your current system (although solving speed is not something that can always be accurately predicted).
On my own current system with 24 threads, this file solves in slightly over 5 seconds.

I'm so confused.. It's late and I haven't been in solverland for way too long.. but..

I wanted to see the difference in EV for a simplified solution so I removed the b80 option for IP after OOP checks. The rest of the tree is the same.
The EV before any action is taken is now higher for OOP as expected.
However, the EV after OOP checks is now higher for IP when his b80 option is removed? I must be wrong but I don't see how this is possible.

Both solutions are solved to 0,1 accuracy. Am I going crazy? What am I missing?

You would need to lock OOP's decision that leads to this decision.
Right now OOP has adjusted his range, meaning that IP is playing versus a different range than before.

Other than that, please check if rake is zero in this tree, or if this is a tournament or a spot with rake.
EV is only maintained between trees for zero-sum spots; non-zero-sum spots have multiple solutions, which makes looking at EV performance far more complex.

If you want us to take a look at the tree, then please send a savefile to support.

In addition to my previous reply, a more straightforward approach would be to leave OOP's decision unlocked.
Then, look at OOP's EV for his very first decision (the one that leads up to IP's decision).
This is OOP's overall EV for his entire strategy.
Mousing over this field will give you IP's overall EV.

Scy Hi!!! , Is there no update coming??

Hello, I'm having trouble changing the color matrix... Does anybody have solid options that are easy on the eyes? Just can't set it to work.

We've needed some time for the new Flopzilla website, which not only required regular content, but also videos and all sorts of other time-demanding work.
That being said, we're basically done with v165 of GTO+ now; it should be released in the near future.

The rake is set to 0%. I also looked at the EV before any actions were taken: there was an EV loss for IP player as expected since he had less options after OOP check. But when I went to that part of the tree the EV was suddenly higher than before for IP. I still don't understand. I didn't save the file as I was just messing around a bit. Will try again and nodelock OOP too (but seems like the EV for IP can only become worse if OOP can deviate prior actions too so that wouldn't solve the mystery).

I use this:

It's difficult to say without the savefile, but most likely OOP had a slightly different betting range in the two scenarios.
As a result, IP's EV went up in the "Bet" line for OOP, and down in the "Check" line.
The overall result will have been that IP's overall EV dropped.
I have checked over here in various scenarios, and did not find any exceptions to this.

That being said, in the upcoming v165, you will no longer need to use the editor to remove IP's "Bet" action.
It will be possible to add a negative incentive to any action that you want to leave out.
So instructing the solver to ignore an action, or to force it to take a certain action will be much more straightforward.

It will also be possible to set a small incentive to an action; for example, to let the solver pretend that the "Check" action performs $1 worse for every hand than is actually the case.
As a result, the solver can be incentivized to play slightly more/less aggressive than would be optimal.

Awesome. I just sent the files via mail. Also with the question how to nodelock OOP bc the strategies reset after changing the tree.. You kinda just answered that spontanously I guess (:
But... It doesn't make sense that OOP loses equity in the check line, since that's where IP is restricted. It also wouldn't make sense for OOP to make more EV in the betline as nothing changes there. Or am I missing something here?

Ok, I have checked with v165.
For this, I have locked OOP's first decision, so that he can't adjust his actions, and forced IP to ignore the "Bet 5.4" action.
When using this setup, IP's EV after OOP's check indeed drops.
From what I can tell, it really looks like, in the two files that you sent, OOP has a slightly different checking range.

Having given it some thought, this actually makes sense.
With IP having fewer options after OOP checks, OOP will be tempted to send some of his worse performing hands to the "Check" line.
As a result, his "Bet" range becomes stronger, and OOP performs better in the "Bet" line.

Equity is not really an important value here, given that it's just a quick, but unreliable method for estimating the value of a hand.
We often use equity in hand analysis because it's an easy value to calculate, but in the end, equity is not a very reliable predictor of the playability of a hand.

indeed the OOP ranges a slightly different, but as OOP can adjust to funnel more hands into a weaker (bc simplified) IP strategy after check, I would assume this to only have a positive influence on OOP EV (in general and when checked). So knowing that the ranges are slightly different doesn't really solve my headache.
Anyways, you seem to have a better setup to find the EV difference after OOP checks.
How did you force IP to ignore the b80?

OOP is checking some of his weaker hands, now that IP has fewer options.
As a result, his "Bet" range becomes stronger.

Basically, OOP had to play the weak hands in one of the lines, and he now moves them from "Bet" to "Check".
The weaker IP plays in the "Check" line, the wider (and weaker) OOP's checking range will be.
It's entirely possible for IP's EV to actually increase locally; the real gain is made with the stronger range that OOP has in other lines.

For this I used the new "incentives" feature in v165.
I expect that we'll release a beta for this tomorrow, so you can check for yourself.
It's a lot easier to make this comparison when locking OOP's strategy.

Oh 😊 Thank you for explaining and for your time! This was very counterintuitive for me but it makes a lot of sense now. I'm looking forward to the new release 😃 I'm planning on diving deep for a while, because I still suck at this game and get complacent way too quickly

Ok, we have posted a beta for v165.
Go here for download: https://www.gtoplus.com/download/

To set an incentive, so that the "Bet 5.4" action is ignored:
1) Navigate to IP's decision after OOP checks
2) Activate "Lock + edit decision", and click on "Special options" (to its right)
3) Select the tab "Incentives"
4) Set a very large negative incentive to "Bet 5.4". -1000 should work, or just any value that's greater than the pot+stacks.

It's probably a good idea to lock OOP's first decision, that leads up to IP's decision.
In this manner, OOP will maintain his range, and IP's local EV should drop after applying the incentives.

Is there anyway to rerun a sim quicker once I nodelock the flop? It takes way too much time for me. I have an 8 core CPU.

Scy, incredible update, AMAZING!. The rounding, incentives, and frequency bar make a fantastic weapon when it comes to nodelocking/building strategies.

I have tried mixing for example "incentives+Rounding"

I don't think this can be done, right?

Another question, could there be the option of using rounding in all runouts style of the "incentives" tool?

Another question about incentives applied to databases:
It is a great tool if it could be applied together.
Let's see if I can explain myself with an example:

In this BD, I want to apply incentives on the flop to choose B25 (range bet) for all of these AHihgs boards.
This would allow the betszings to be adjusted on the flop without losing the agg reports as a whole.

Better yet, in AKx have 2 strategies of B25 B75 and 0 checks.
They are examples.

In conclusion:
- be able to apply incentives from the flop in DB mode on several flops at the same time.

Once changes have been made, the solver needs to again figure out the optimal play with the given restrictions.
There's unfortunately no shortcuts for this.

This is fairly straightforward from a technical standpoint, but the interface may be difficult to work with for this.
So, we'll consider it for future releases, but I don't immediately see how we would accomplish this.

It's not possible to mix incentives with rounding, given that incentives require the solve to be run on an unlocked decision, whereas rounding is a form of locking.