Ah, I see the problem. The summary box below the table actually gives a summary of what is currently in the table. However, when applied to a specific action it's probably better to just keep displaying the summary for the entire decision. We already do this in the navigator, but it's probably a good idea to do this in the analysis tool as well.

Thank you for pointing that out,

Scylla

No problem,I have noticed that only entire decision summary gives the clear picture.Maybe u can consider ,when displaying specific action,hands that not fall in mixed strategy(100% or 0% hands) to be shown separately and not shown in both samples.It will give clear picture.(like in pokersnowie range advice if you have seen it).

Ok, I've received your savefile. Basically, equity is not as telling as a quantity as people often think. It's very frequently used in books, videos, etc because it's relatively easy to calculate, but the problem is that on the flop you don't get to immediately take your hand to showdown. After the flop there's still the turn and river to come.

See the screenshot below of EV versus equity in a random postflop spot. The blue line represents your expectation based on equity. As you can see, equity is actually a pretty poor predictor of EV. Even two hands with identical equity will often have a different EV, due to implied odds and reverse implied odds. And reverse implied odds are where this goes wrong for the OOP player in this hand.



Hands that have great implied odds are hands like flushdraws and straight draws. Technically, based on equity, you may appear to be a in a weak spot. However, when you hit your hand, you stand to make additional money. Typically, flushdraws and straight draws will perform above the blue line in the above pic. So they stand to make more than you'd expect based on equity.

On the reverse of that are hands that suffer from reverse implied odds. These are typically one pair hands with no real draws. For example, a hand like AJ on JT9 is a hand that will often perform below what you would expect based on its equity. With a hand like AJ, you have no idea where you are at in the hand. You may already be behind, villain may hit a straight or two pair that beats you, or a scare card (KQ87) might come off where villain can force you to fold.

The range of TT-77 is similar in this spot. You are either way ahead of villain's 66-33 or way behind to AA-QQ. And villain is in position for all three streets. It's not possible for you to play aggressively, given that in that case villain can make full use of his AA-QQ range, while folding at least some of his 66-33 hands.

The consequence is having to play a passive check-call game out of position for three streets versus villain. Given that there's many ways for the board to run out it's difficult to figure out exactly what the breakdown here is, but apparently villain is in a position to make you fold on the river in case he has the option to make a huge bet.

Basically, the play will go check-bet-call, check-bet-call, and again, on the river, check, after which villain will bet all of his AA-QQ hand, mixed with 66-33 in such a frequency that you can not call profitably enough to justify your earlier calls on the flop and turn. For all possible runouts of the board, there's just too few scenarios where you hit something.

Yes, v106 will indeed contain the preflop ranges. The preflop menu was the first menu that we originally designed. At that point in time it was not yet fully clear how the rest of the interface would end up looking. Building an interface is something of a creative process, and it's not always entirely clear what the end product will look like. As a result, particularly in the earlier menus, we have left out quite a bit of detail so that we could fill that in in the final iterations. So, given that the preflop menu was the first menu that was designed, it's also the menu where most details have been left out. Other examples are the "Build tree" menu and "Run solver" menu, where originally we only provided the "Basic" tab and left out the "Advanced" functionality.

This will indeed be possible. Currently the Flopzilla and CREV format use a newdefs2.txt file. For GTO+ we will probably move to newdefs3.txt. So all you will need to do is copy your newdefs2.txt file to GTO+ and it will convert it to the new format.

When you are ready for the trivial changes ....Making this clickable would be spectacular:


Thanks

Yes, that's probably not a bad idea.
Thank you for suggesting that.

After v105 we are now offering an advanced tree builder and tree editing options. I don't think that there is much of a justification anymore. In fact, we even offer many features that other softwares either don’t offer, or ours are far more user-friendly. As for solving speed, this is similar to other products (sometimes a bit faster, sometimes a bit slower). I will just create a summary of what I feel are the most important differences. See below for that.

Other than the points below though, you can also take a look at our release videos so far: http://www.cardrunnersev.com/gtoplus.html. When watching the videos, ask yourself if those other tools offer these features. I think you will find that they either can't, or that they offer an alternative that is far more difficult to work with.



Internal analysis tools
First of all, we offer internal analysis tools whereas other tools require you to manually copy-paste ranges to other software in order to see the composition of the range. I feel that it’s just a very inconvenient way of looking at a solution, and the constant delays due to having to manually copy-paste will really get in the way of your learning curve.

Graphing tools
We offer internal graphing tools to plot basically anything versus anything. We also offer these same tools to visualize turn and river reports and aggregate reports. Other tools offer nothing other than text based output, which, again, requires copying to external programs. Most data is lost in translation here, as well as the delays getting in the way of your learning curve. You're also far less likely to run into anything by chance, given that you only get data if you have specifically requested it.

Tree navigator
Other softwares only provide line-based navigation, whereas we also offer a full extensive view of your tree.
To get to the navigator, just press TAB.
Mouse over any part of the 2D tree to get more info.

Aggregate reports
We offer aggregate reports with internal analysis tools, as well as several additional output options.

Tiny savefiles+databases
Not only do we offer the ability to store to very small savefiles (roughly 100kb for a tree), but also the ability to store multiple trees into a database, with the ability to make comparisons between those trees and to create aggregate reports. Where typically solving that many trees would take roughly 10GB of storage space, due to our method of using small savefiles, in our case this will take roughly 10MB, with no data being lost.

Recalc options
In GTO+ you can easily recalculate any turn/river subtree without needing to resolve the entire tree.

GTO+ converges to 0%
Where the other softwares often stop converging at around 0.05%, we converge all the way down to 0%.

Tree editor
We offer far more convenient methods for making edits in your trees with a full view of the entire tree. I believe in for example pio this requires text-based input.

Process all files in a directory
We offer the ability to process all files in a given directory. So, for example, you can create multiple databases of many trees that you would like solved, put them all in a directory, point GTO+ to that directory and let it solve all files in it.

Card removal
We offer an option to toggle card-removal on/off so you can easily see the effects of card removal at the push of a button

Decision editing + locking is far easier
We offer visual tools for directly editing your decisions, locking them, and re-running the solver.

Our main disadvantage is that GTO+ is relatively new and not that well known yet. So people considering a GTO solver are less likely to run into us. With us leaving beta we hope that this will change in 2018.

Hello,
thank you again for your work.
I have CREV, flopzilla, and I am the evolution of GTO +.
But when I connect to the site I see only version 1.04 available.
Am I the only one.
Thanks for your help

Scylla

I'm completely new to this and haven't used Cardrunners EV (I do have flopzilla though ). So I've purchased it yesterday for GTO+ but as a noob it's a bit daunting. I've watched the 3 youtube videos and started playing around with the software to familiarise myself with it.

If I was looking at revamping my whole BTN v a caller in the blinds strat (I'm thinking of one where I'm range betting 1/3 on a lot of flops so it's the turn etc decisions I'm most interested in) what would be the best way of using GTO+ to come up with an overall strategy? I'm not looking for detailed instructions just some general pointers as I'm trying to avoid slogging away looking individually at various boards only to discover there's a shortcut or easier way of doing it.

Thanks

Generally you can create a tree and solve it. After that, you may want to set a strategy for villain in certain decisions in the tree. It doesn't need to be for all decisions; just the ones you're interested in. To edit a decision, click on "Lock+Edit decision" (lower left) and edit villain's decision in whichever way you see fit. After that, re-run the solver to see how you should adapt to villain's strategy.

Should you be looking at the turn for a flop hand then it's not necessary to resolve the entire tree. You can also just re-solve that particular turn by clicking on the "re-solve" symbol; in this case the re-solve will only take a few seconds.

It sounds like you're loading from cache.
Just press F5 (or reload) to refresh the page.

Scylla

By what formula is the check-raise calculation % pot ?

Thanks

Okay, thanks

The pot is 5.9.
Villain bets 2.95.
After you have called the bet, the pot becomes 5.9 + 2.95*2 = 11.8.
40% of 11.8 is 4.72, which will be your raise on top of the current bet of 2.95.
4.72 + 2.95 = 7.67.

number 2 is the constant value in the formula?


The pot is 5.9.
Villain bets 4.43.
After you have called the bet, the pot becomes 5.9 + 4.43*2 = 14.76.
60% of 14.76 is 8.86, which will be your raise on top of the current bet of 4.43.
8.86 + 4.43 = 13.29.

What kind of computer specs would you recommend to use GTO+ completely fluently, so without any freezes, substantial loading times etc.
Do i have to expect any issues with 16GB of RAM? What kind of RAM usage is the Maximum you encountered using GTO+ yet? Would 32GB or more be benficial using the program?

The starting pot is 5.9.
Villain bets 4.43.
You call 4.43.
So 4.43 goes into the pot twice


Yes, however, you have the option "With only 2 bets left, get the money in smoothly" turned ON. So the players will make slightly different bets than instructed in order to size the two final bets more smoothly. This reduces the size of the tree, as well as making the play more realistically.

With this option OFF players would either make their final bets too small, leading to them having a tiny amount of money left in their stack, or too large, in which case they are risking more money with their bet than they need to, as well as committing themselves.

16GB is totally fine. For example, see the pic below. This is a profile for a tree with 3 different bet sizes for both players. It requires only 6.7GB. However, due to compression, which is automatically applied when memory use exceeds 3GB (the performance penalty is 10%), you will actually only need 3.35GB.



Please see below for a pic of what this tree looks like on the flop. As you can see, this tree is huge. And it will look similar for all possible runouts on the turn and river. Although I wouldn't tell anyone how to perform analysis, to me personally, this tree already way exceeds any size where there's still something to be humanly learned from it. As I show in the latest release video, sticking to only a single bet size does not seem to carry any significant performance penalty. When performing analysis, even with multiple bet sizes, I myself would not bother with too much complexity beyond the point of the tree that I'm interested in.



For example, when using a slightly smaller profile, with only default play on the turn and river followup play, which should lead to similar results, the size of the tree is reduced to as little as 2.9GB. This will make it solve a lot faster, while the quality of the results should not be affected.



You may notice that other softwares have far higher memory requirements. This is due to them currently not offering the option "With only 2 bets left, get the money in smoothly". People sometimes solve this by just spamming the final bets with multiple bet sizes, in the hopes of hitting a sequence that makes sense. With our approach, this is not necessary.

Thanks a lot for your answer and your wokr here in the thread. You got yourself a customer

Hi,

To follow up on the question about pc requirements, what about the processor? What would you recommend here for everyday use?

For example, i5 vs i7 vs Ryzen 7 (I know it's gonna depend on the number of cores and generation of the CPU but some general observations if you tested various processors.

And would you even recommend using it on a laptop with i5 (2 cores) with 8 gb or is it better in that case to rent a server?

In the use custom settings for (flop/turn/river) section of the advanced tree editor. Could you potentially add the ability to do different variables alongside % of the pot.

Like instead of using 50% of pot to make a 50 chip bet into a 100 chip pot, we could put 50c in the box. Or in the raise boxes, we could put something like 3x to denote that our raise size will always be 3x of the opponents bet size.

Ok, I'll take it into consideration.

Thank you for the feedback,

Scylla

I would recommend a CPU with at least 4 cores. The speed of the solver will basically be determined by the CPU speed multiplied by the number of cores. So a 4 core 3.2gHz CPU will come down to running on a 4x3.2=12.8gHz system. The memory will not affect the speed; it will only affect the largest size tree that can be stored in memory. As discussed in post #6469, GTO+'s memory use is relatively low, as well as it being able to apply compression. You should be able to run some pretty large trees with 8GB, even if not all of it is freely available. For example, the tree discussed in pot #6469 with 3 bet sizes for both players throughout the entire tree only required 3.35GB after compression.

+1