Is there a way to display equity and strategy at the same time in the 13x13 grid? I know we can do it with EV, but equity would be helpful too.

Ah ok thanks - it was two separate posts by the way, you must have missed the first one.

Stupid mistake on my part - I'm now worried that lots of the sims I've run have had crazy results because IP never had the option to 3bet smaller than all in!

Any rough guesses as to how much of a difference this would have made to 100bb deep SRP study that I've been doing?

There is no way to do that now but I will put it down as a suggestion.

Probably not that much difference but the only way to know for sure is to re-run a few trees and compare the solutions.

When I run a sim I like to use multiple bet sizes and pick the best one. If I plan on using 1 bet size in game should I make some assumptions?

For example in 3bet pots if I put 4 bet sizes 33%,50%,66%,100% and it bets 65% of the time, checks 35% of the time.

Running this same sim again with only one 33% sizing it bets 81% of the time and checks 19% of the time.

In the future should I just make an estimation when using multiple sim sizes that I can bet 10-15% more if I'm only sizing 1/3 and a bit less if I'm only sizing 2/3?

This makes sense. In general if bigger bet sizes are available and it chooses them with some frequency the overall betting frequency decreases. The general principle is: the more you bet the less frequently you should do it.

Running a sim with many bet sizes and then choosing one based on this isn't mathematically guaranteed to work but it seems to be a very good heuristic. It's hard to say if 10-15% more rule work in general though. It's the best to re-run the tree with one bet size. It will be much smaller so it will only add small % to the total calculation time.

Cool that makes sense I'm actually going to remove the pot size bet OOP to make it more accurate then. I never use that sizing there.

Hey guys,

I have bought PIO but I want to optimize my PC so I can solve more taxing stuff.

I have understood that it’s the processor and RAM that is the most important.

I am currently sitting on:

Intel Core i7 6700K 4.0 GHz 8MB
Corsair 16GB (2x8GB) DDR4 2133MHz CL13 Vengeance

Will it be enough to just upgrade my RAM to 64gbs or will the processor be too slow also? Also does it make any difference what type of RAM I get?

I am no computer guy, so I have no idea about this stuff.

Is there a way to customize aggregation reports? For example, can I pull EV or Equity information not just for the entire range but for a specific hand or like all top pairs?

I have Just purchased the Basic version, I am a long term player so understand poker fundamentals, I have a few quick questions about PIO.

What are the most important things to look for when seeing what hands PIO chooses for each decision.

If you where beginning today with PIO do you have any advice, tips, tricks?

Hi, how can I run multiple consecutive scripts without having to start it manually again?

More RAM lets you build bigger trees, faster CPU lets you solve them faster.
CPU speed is proportional to number of physical cores multiplied by base frequency they run on so you can estimate how much faster other CPUs would be.

This is not possible at the moment. You would need to filter those hands yourself.


I think range composition is the most important. For example it's useful too look at the betting range in range explorer and see what % of that are strong hands/draws/weak hands and which hands in each category are chosen more often.
This is also something which has the biggest potential to change your current game. For example many years ago when I was playing people in general had too strong betting ranges and too weak checking ranges.

It's not very convenient right now. See post #5214 in this thread:
https://forumserver.twoplustwo.com/s...postcount=5214

Hi, does anyone know how can I use big blinds instead of chips in pio?

Would it be possible to make range explorer a little more robust, with a few more categories? This is the way GTO+ does it:

You can't but you can just use 1bb = 100 chips and you will have results in 0.01 of bb.

It looks like what Pio shows, what exact information would you like to see which isn't available now?

I'm working on a solver and wondering if you could give me any hints. Are you storing strategy probabilities as 10 bits each? This allows three decimal places. And then I'm assuming regret sums and strategy sums are stored on disk and you're compressing/decompressing on the fly, unless you're not using cfr at all.

Just saw you're not using CFR.

We don't use CFR, we use something that uses way less RAM. Our algorithm has advantages (way less RAM, fast convergence to something decent on trees with many bet sizes, doesn't jump around as much as CFR+ so it gets overall frequencies in the correct ballpark faster) and disadvantages (has problems zeroing actions so it often ends up with something like 0.997-0.003 instead of 1.0 - 0.0, usually stalls somewhere around 0.25%-0.1% of the pot).
We might ship CFR as an option in the future although right now the most important things our users want is to be able to build bigger trees and for that you need to save RAM.

Hi, I am going abroad for several months and during my time away would like to utilize my pc and run pio sims while I am gone.
For my use case, which do you think would be better (I don't need immediate access to the solutions fwiw):
1.create a chain of scripts (the .bat file method) and simply let it run
2.utilize a remote access app like teamviewer and control pc/execute sims from a laptop abroad

Any input would be dearly appreciated, thank you

2nd option, especially if you can ensure your PC stays on (maybe you have someone to turn it on if something happens). This way you can modify what's being done on the fly should you change your mind or have some new things to run.

Is there a way to make it show what the strategy is with each category? The colors on GTO+ represent bets and checks.

Sure, you can select one or more categories and the strategies on 13x13 view will update. This way you can show strategies for top pair, top pair or better, top pair or flush draw or any other combination like that.

Thank you very much punter, will go for the second option

Hello, I have a question please :

When scripting I solve for 0.5% pot but I always round everything to 1/4 and the end result is exploitable between 1.5% and 5% (which I'm fine with). In my case if I solved for 1.5% pot instead of 0.5% would that give me the same end result?

EDIT : And would that work in any given situation, not just 1.5%? Like in a situation with wide ranges where rounding to 1/4 results in being constantly exploitable for 5% would solving for 5% give the same end result as well?

Thank you for your time.

why is there a range explorer? Why cant that same functionality just be built into the main browser?

It's not guaranteed mathematically. My intuition is that solving to 5% might yield still quite bad results but if your intention is to round anyway then likely anything <1% is fine.

The reason is screen real estate. There are many buttons and features in the Range Explorer and there was no way to fit them all on the main view.