Hi there,

I was always a bit bothered by the recalculation time (with basic storage) for the turn, I like to play fast against the solution. I thought a better processor would help, but it doesn't: Is there any explanation, why my 8 GIG RAM / 4 core intel laptop is significantly faster then my 64 GIG RAM/16 core AMD Epyc Server? I am really puzzled

This is not true for the solves however. Here the server is as expected significantly faster.

also: Can I convert a solution from basic to extensive storage? It seems that I would have to recalculate everything from scratch. correct?

cheers!

I have a database saved with medium storage (turn reports for individual hands).

When I go to select an individual hand in the turn report section, the moment I select the 2nd hole card, the program closes and throws me out.

Is this a known issue? Could it be something on my end or a bug?

Yes, this has been reported. I haven't yet had the chance to look into it, but with "Medium" storage no longer being available in newer releases, it appears that we may have overlooked testing for this scenario. I will take a look for v131.

I can't really think of any reason why your laptop would be the faster machine here. I'd have to check the code, but I'm fairly certain that all threads are being used. As for converting the storage method, the turn data is not available in "Basic" storage, and can only be determined from a new calculation.

Hi scylla,
Is it possible to live export a range from GTO+ to flopzilla by decision? for example if I ran a sim and want to check my equity against villain's double barrel calling range can I export to flopzilla just the calling range from his whole range?

For this, go to the tree navigator, go to villain's decision and select the "Call" tab. You're now looking at your range versus villain's calling range. Press Ctrl+F to export to Flopzilla. It's however not possible to do this from within the "Play against the solution" feature.

Some info for your investigation: I am using Version 1.2.8, on the laptop it was installed with the installer, while on the server I use the zip distribution. I also did a clean restart of the server with same result: the laptop ist significantly (~25%-33%) faster.

Have you actually set the correct number of threads for the server (under "Run solver")?
If you're using too few threads, then the server will run slowly.
Does the server hit 100% CPU usage while solving?

Any chance you can implement copying and pasting ranges for a specific color in the future? For example, on one particular site, I'm able to copy a range and paste in green. Then I can copy another range and paste it in blue (there will be errors of course if the combined range equals over 100% for a hand). This would be much more convenient rather than combining green and blue and making the ratio 57/43 and then clicking another hand and adjusting the ratio to 62/38 and so forth.

Ok, I will consider it for future releases.

Or maybe show hand weight as a percentage for each combo

Scylla,

Is there any built in software protection of calculation errors from say an unknowingly "unstable" CPU overclock? Not unstable in the sense that the system will shut down or freeze up but from what I understand CPU's can make miscalculations as well and I would consider that to be unstable. Will the program crash or be able to detect and resolve a miscalculation, or is miscalculations not even a thing? You input is much appreciated.

I'm currently running an all-core overclock on a 3900x to 4.35GHz and my system runs stable in back-to-back runs in Cinebench R20. It doesn't it crash during several multi-hour database calculations in GTO+. I can get 4.375 stable with some voltage increase, but I'm currently unable to get 4.4GHz all-core stable. So, you see, I'm getting as close as I feel comfortable getting to the line of stability here, but so it makes me question how safe does this sound for GTO calculations, in your experience?

would be nice to set the cards in this options.

https://prntscr.com/taisvn

I would also like to humbly ask if we might be able to get a "Clear" button in the range editor-- to zero out the entire range for that particular position being observed. I've been meaning to ask for a while because its a common reoccurring delay I encounter, especially when working with iterations. I only ask because Flopzilla Pro has this feature for its range editor.

Hi,
I'm contemplating between GTO+ and Pio Solver. I'm aware of the price difference, but i'd like to know the difference in function. I'm looking for the most extensive option. Thank you very much.

For brevity, I will just copy paste an answer to a similar previous question.

Solving speeds
Our solving speeds are similar to pio.

Play against the solution
We offer a "Play against the solution" feature, which is not available in pio (here is a short video: https://youtu.be/YuRsCQajRBA).

We converge to 0%
GTO+ converges to 0%, whereas pio will typically stop converging at about 0.05%.

Small savefiles and database functionality
We offer small savefiles, requiring only a few hundred kb per save (non-stored data can quickly be recalculated on the fly). Our small savefile approach allows us to let you create databases of hundreds of trees, which is something I believe pio can not offer.

Internal analysis tools
We offer internal analysis tools analysing the solutions, as well as graphs, tables, etc; to the best of my knowledge pio only offers the most basic of analysis methods, while otherwise needing its output to be manually copied to external software.

Visual editor as oposed to text based editor
We offer an editor with a graphical interface for editing trees, whereas pio only seems to offer some sort of text based editor.

Card removal ON/OFF
We offer a feature to toggle card removal ON/OFF throughout the solution.

Multiple bet sizes
We do indeed have multiple bet sizes (see video 3 on the tree builder here: www.gtoplus.com/videos).

Other differences
The only thing that we don't offer is a preflop solver. It's possible for us to create this, but the problem is that this feature requires a custom built computer, which almost none of our target audience will have. With some searching you may find some minor differences between pio and GTO+. We could easily add whichever features those may be; however, the decision not to clutter our interface with all sorts of buttons is more of an interface design choice, where we need to balance how many features are offered in the available space versus ease-of-use.

Why is the price lower?
GTO+ can essentially do everything that pio can, and is even in many ways superior. Some people assume that because we offer it for less, there must be a reason for this. However, for example, our other software Flopzilla is also sold way below market value, with lifetime upgrades for only a single payment of $25. The reason for the pricing is simply that we don't want to charge regular consumers huge amounts of money for gaming-related software.

It's probably theoretically possible for a CPU to miscalculate, but within use for the solver, it would at worst mean that the solution would not converge. If the solver calculates a certain dEV, then this dEV will be the correct value, even if due to some internal workings one or two hands would be incorrectly assigned to a certain action. I suppose that miscalculations are a more important consideration for critical systems, such as space shuttles, military equipment, etc.

Your CPU has been provided with certain specifications. Outside of those specifications it will indeed most likely function, although at the very least the producer of the chips is not willing to offer any guarantees.

You can set a specific hand for hero under "Lock hand for hero".

A "Clear" button has been added for v131, so please check back for that.

Hi,
I just wanted to know if when we play against the solution the "RNG" number that appears would change in someway the outcome. e.g. with RNG: 15 call 80% and with RNG:100 call 100%

I was wondering if it is possible to see an agregattion repport for the turn, the same way is possible to do it for the flop.
https://gyazo.com/454a636b80a946af6f3d4730d68100d9

What is the progress bar measuring exactly?

Thanks a lot.
As far as a preflop solver, do you have any software to accomodate GTO+ in that regard?

Any discount for the purchase of Flopzilla if I buy GTO+?

Any plans to have GTO+ for PLO?

So basically, GTO+ does every calculation that Pio does in a more intuitive way? recommended betting sizes based on EV calculations are available as well?

There's many different sources for preflop ranges. Both ranges based on experience, as well as ranges based on analysis. We're not particularly invested in designing GTO software specifically for this, given that almost no one has the hardware to run it anyhow. We prefer to focus our efforts on features that are within reach of all of our users.

Flopzilla costs $25 for a lifetime license, with all upgrades included. For example, even people who purchased in 2009 still get FlopzillaPro for free. We don't offer additional discounts on top of that.

Unfortunately, it's not possible for me to discuss development plans.


I would say so, although I am of course not an independent source of information.

In GTO, all bet sizes will perform really closely to one another. You can test this for yourself by building the same tree, but for different bet sizes. After that, check OOP's overall performance (his EV below the table in his very first decision; see the screenshot below). So, it's not possible to recommend bet sizing, given that all sizes will perform almost exactly the same when playing perfectly. As a consequence, we don't offer such a feature, and neither does pio.

What some people do here is use multiple bet sizes to see which action is used most frequently (it's possible to do this in GTO+). However, this approach doesn't really work, because the performance of the bet sizes is nearly identical. Whichever bet size is used most frequently is mostly just a coincidence, and different solving algorithms will find different "preferred" sizes.

The RNG number is just a randomly drawn value from 0-100.
It does not affect any calculations.