Another question.. when using multiple bet sizes is it possible to quickly change to see how the strategy for both players would change if instead only one sizing was used? For example, perhaps on a certain texture the solver prefers something like betting 2/3 pot 50% of the time, 1/3 pot the remaining 15% of the time we bet, and checking 35%. To simplify our strategy we could just decide to adopt the larger sizing with all of our betting range instead of trying to figure out how to properly balance 3 different decisions. Is there a way to do that without having to solve the entire tree again? I saw somebody do this in a video using PIO using the node-lock feature which of course exists on GTO+ but I'm not sure if that specific ability exists. If not, I would love to see it added (and at a much higher priority than my last request).

Anybody care to share some CREV files? Like some interesting spots you have analysed. Im going through the online videos and manual and Im looking to learn and see how some of you guys went about building trees. If anyone feels like sending me some Tree's you've built in CREV or Spots in GTO+ please send me an email at PLAYTOLIVE81@HOTMAIL.COM . I would greatly appreciate the help. Im struggling a bit with this software.

Thanks

here is example of crev analysis I recently sent to the forum.

Thanks for that! I just skimmed over it, so it looks like a tourney spot where BB is the shorter of the stacks. I noticed on the flop you setup actions starting with BB etc. I guess you dont need to setup an actual flop in order for it to continue flop calculations? I notice the flop cards where still blanked out and not chosen?

thx

You can leave flop empty so that crev would be able to run number of simulations(you can see this number on the top of the tree, it's 100K). In this specific spot I wanted to check the EV of calling with marginal hands from BB and shoving if certain combinations were hit OTF...

Simplifying strategy is not easy if solver recomends 50% line.U will have to sacrifice big % of Ev to do that.But if u create a model that prefers one bet(small bet usually 1/3 of pot or smaller)than in many spots solver will bet 75% or more on the flop as continuation bet.In this spots with lock mode u can simplify strategy that will bet 90-100% with a loss of EV of 1-2%(u will lock the weakest combos in your range to check 100% of the time so for u to bet with your whole range close to 100% and still stay balanced).

Ok yes I understand. Im trying to come up a proper system for analyzing hand histories or theoretical spots, doing some off table analysis. Im thinking running the hand through flopzilla and equilab first, trying to reach some conclusion's and then finally analyzing the hand with CREV/GTO+, to see if my conclusion's where correct? Its tough if you're new to this kind of software feels a little overwhelming. I kind of have to find a process/routine to follow every time that will yield results.

If you're new to the software, then I would recommend using GTO+. It covers many of CREV's features and has a smoother learning curve. GTO+ is our latest software, and although not all of CREV's features are covered, it can to a large extent be considered to be CREV2 (focussing on the GTO aspect).

Can you please take a look at post #6823 a few pages back? This subject is discussed under "The importance of bet sizing" at the bottom. Here is a link to the savefiles: www.crevfiles.com/crev/forum/1/processed.rar. As it turns out, bet sizing is almost irrelevant. For just about any bet size a GTO solution is possible that performs about equally well as any other bet size. Another point of interest is that using multiple bet sizes does not give any significant advantage over only using single bet sizes. This is discussed at the first part of the post.

You can do this by editing your tree to add the requested changes to the bets. After that, you will need to solve it again though. In all likelihood you will find that the player's overall performance does not change significantly when using just a single bet size.

I definitely agree with this, but I do think exploitatively using multiple bet sizes in some situations can yield additional EV because humans have a difficult time defending properly to things like small flop cbets and turn / river overbets. Because of that I think it's useful to explore when the solver might employ multiple bet sizes with similar frequencies if given the option and when it strongly prefers to opt for a particular sizing.

In any case, I think one of the ways to snuff out what the optimal sizing may be in a particular situation is to give it various options and see which it prefers. From there, we could decide to edit the tree and opt for the one betsize that it was using most frequently and run the solver again to see how each strategy changes. For example if we use a 3/4 pot sizing and a 1.25x overbet on the turn, it's basically always going to bet choosing to overbet things like overpairs and TPTK if that sizing has any frequency, while opting for a more traditional size with the weaker portion of our value range. It seems reasonable to think that our opponent could defend differently to the 3/4 pot bet if he knows we're only using one sizing because blockers to the very top of our range are now more valuable when we're not using them to overbet. The inverse of that is that blockers to the top of your opponents range are far less valuable when they opt for a 3/4 sizing in a tree where they are also overbetting with a reasonable frequency.

I'm just getting into the world of solvers after a fairly long hiatus from poker so I'm still trying to figure out the best use of the software. I'm finding immense value in it so far though and I really appreciate your frequent replies in this thread.

That's kind of what I figured. Am I wrong in claiming that PIO is able to do this without having to resolve the entire tree? I believe I saw it in a RIO video but it's possible there was some slick editing in play that I missed.

Hi scylla

Im encountering an issue. When I hover over Edit conditions in CREV it says to left click to edit conditions and right click to multiconditions menu. When I left click it opens the hand range dialog box. When I double click it opens the hand range box and when i right click it opens the multiconditions box. I cant open the edit conditions box it automatically opens the hand range box???? Ive tried everything and resetting thew program etc.

@Scylla, Never mind I figured it out. I was doing it all wrong. thx

@scylla besides the fold all hands action being at the bottom does it matter if raise/bet or call order of actions in CREV? Im trying to figure out if we should follow the same order all the time e.g. Raise, Call then Fold?

Thanks

Also, im looking for a document or part of the manual where it says how to interpret the EV results. Is the green EV number a sum of the EV results of each action in that particular node?

Im asuming the EV calculations at the top of the screen is a sum of the EV numbers of the entire branch.....

It may indeed offer an advantage to use multiple bet sizes against human players, even though from mathematical point of view it does not really seem to matter. Human players are far from perfect, so anything you can do to make them play even further away from perfect is something to consider.

It's either that or by building several different trees, each for a different bet size (like in the post that I linked to earlier). The latter method is more accurate, although the most interesting thing to learn from it is that there's apparently a similarly performing GTO strategy for just about any bet size.


It tends to play all hands in a mix, so that both lines have strong hands in them. If it were to assign nut hands to only one particular line, then villain would know you're weak in the other one.

I'm not sure if it offers that. We can consider it for later releases, but right at this moment it's not there.

I would recommend going from strong to weak, so bet>call>fold.
For GTO+ we have gone so far as to always use this order.

Yes, it's the weighted sum of those EVs.
For example, let's say that a decision has two actions.
A bet for an EV of 26 and a check for an EV of 10.
The bet is performed 40% of the time, and the check 60%.
That makes the overall EV for the decision 40%*26+60%*10=16.40.

When we solve a spot,on the flop and we look for results(equty and EV of a range)we can look only for player's 1.For player 2 we can look only after player 1 has taken an action.Why is that or am I missing something?We should be able to shift in both ranges to see who's range has advantage before decision is made(to see who should be pushing equty)because after that equty and EV r different.

Another question,what is the difference between basic,medium and large storage when solving?Can we take some advantage of that?

This is for the reason that player 1 is the one who is actually making a decision.
At the very least for the moment we have chosen to only display the data that's relevant to the decision making process.

If player 1 has an equity of 80%, then player 2 will have an equity of 100%-80%=20%.

Basic:
The savefiles will be very small. Turn and river data is not stored, but recalculated whenever the user requests it.

Medium:
Same as basic, however, now, in the turn reports, data for individual hands is stored as well.

Large:
Savefiles will be larger, but turn data is now stored as well. This means that, as opposed to basic storage, it does not need to be recalculated whenever a turn is requested by the user. Recalculation typically only takes a few seconds though, and in most cases the difference with basic storage will not really be noticeable.


Overall, basic storage is the recommended option.

Yes,it is obvious for equty but there can be significant gap for EV calculation because of nutted distribution.EV calculation is not proportional.Anyway,thanks.Usually we make decisions based on equity advantage.

Oh, sorry, I overlooked EV.
When it comes to EV, we have the equation EV1 + EV2 = pot.
So the sum of the EV for player 1 and 2 will be the pot.
Therefore EV2 = pot - EV1.

Hi scylla

In database mode, can I use nodelocking by one click?

For example, if I want to see the results when I c- bet half-pot size in 100%

frequency in fifty different flops, should I set every flop one by one?

I want to know if there's any method to set nodelock for several flops , in one

click.

It's probably easiest to just edit your tree so that you only have one action where you c-bet half-pot. And then create a database from that tree.