Yeah, many of them even less than that. I am surprised the postflop accuracies are that big at 2-3bb tbh. I will take a look myself once I am testing some preflop trees.

Great, thanks

PioCloud's pre-packaged solutions use 60~80 flop subsets. Do you agree with him that it's a waste of time to go higher than this for 100bb? How much greater accuracy can we expect to attain going higher?

I think solutions on those subsets should already be very good. I am not sure about "waste of time", it depends how important it is for you to get details exactly right and how much budget/time you have. I think they were mainly 100-150GB trees which already take around about a day to solve to accuracies PioCloud does even on a very decent server so if one can add say 30-40 flops at the cost of more expensive server and additional 8-12 hours then yeah, I can see how it's possible to call it "waste".

There are a lot of trees in a pack as well so raising subset size by say 50% could result in costs being 100% higher (you need both 50% more time and more expensive servers).

Hi!

Is there any way, to export data from range explorer? (Like TP bet x = 75% check = 25%)
Like the aggregated freq analysis

Not at the moment but adding frequencies by hand category is on the to-do.

When can u add this feature approximately?

It would be great!

I don't know, we don't do ETA's or timelines. It's not top-top priority but it's something often requested and reasonably close up there on the to-do.

I tried doing a more accurate calc w/ the same configuration as before at 25bb and the flop exploitability went down to 1%, so I guess even though preflop ranges don't change much flops do get quite a bit more accurate

That makes sense as EVs/general frequencies converge way quicker than strategies for particular combos (that's why it's already quite close when it comes to overall picture even at higher exploitability). Still 1% at 2-3bb/100 sounds too big to me. I believe you are doing it correctly but if you don't mind sending me a small save (preflop+flop) somehow (maybe upload to dropbox/google driver and pm/email a link) I would like to see it myself

It's actually 1% at 1.5bb/100, but yeah I'll send it to you over Skype

If anyone is wondering it's solved. It's actually ~0.4%. There was a slight misunderstanding about where total pot is displayed.

Hi,
just wanted to ask how many threads the edge version supports

It's 128 threads right now but if you have hardware requiring more, please contact us and I will make a compile for you supporting more.

Thank you for the response.

Anyways I bought the software and am very happy with it. It's both very impressive and easy to use.

It would be nice if it were possible to create a subtree configuration on the flop from a preflop solution file for when you want to calculate a flop that wasn't included in flop subset.

Yes, added to the to-do.
What you can do for now :

1)Navigate to the preflop exit you want in the preflop config:
https://gyazo.com/cab25f7160cd31111602e16629ae5936

2)copy the configuration there:
https://gyazo.com/6b1c7396b80ba2f961bcecea73ad57cf

3)go to Tools->paste tree building config from clipboard

You will have the config with betting structure, starting pot and stacks but sadly not ranges. Those you need to get from the solution by navigating to the flop exit, showing range there and copying it (by right clicking on 13x13 area):

http://imgur.com/a/6TSyH

It requires some clicking around but still better than inputting everything by hand.

There shouldn't be any problem with running multiple sims at the same time on one computer provided I have enough memory, right?

There shouldn't be although:

-if you run too many it might be detrimental for performance
-it's a good idea to make copies of the Pio folder and start Viewers from those separate folders. The Viewer is not designed to be run mulitiple times from the same folder as it has only one settings file so if you you have multiple ones and start changing settings you may run into errors

Okay and is there any way to see the tree's file size before saving it?

There isn't but good approximations are:

-full tree save (not recommended) is about as big as the tree in RAM
-small saves (flop + turn) are about 75x-100x smaller than full save
-micro save (flop only) is about 1000x smaller than full save

For browsing small saves are better than full ones because they load much faster and recalculating rivers (which happens automatically) is very fast. You won't be able to change strategies (lock nodes etc.) on those trees though and running aggregation reports on rivers is going to take a long time. I talk a bit more about micro/small/full saves, here:

https://www.youtube.com/watch?v=VhG_YsEvfwU (question number 5)

Question about computer recommendations for using PiO comfortably? I have now i5 3517 wit 12 gb ram and solving takes quite long, 30 minutes or so for 2gb trees.

I´m upgrading my laptop anyway and would like to know if laptop with i7 6700HQ and 8gb ram would be enough to solve trees much faster.

This is a typical laptop CPU with only 2 cores clocked at 1.9Ghz so it's expected to be quite slow. 2GB trees are bigger than they used to be though (because of memory optimizations in recent versions)

I would expect it to be about 3x faster than your current CPU.

How are the flop weights used when running an aggregated analysis over multiple files? What exactly does the number mean when you see flop:1.5, the 1.5 is not a probability because its greater than 1. So what is it?

What does global % mean in aggregated analysis file column?

Thanks.

Weights are relative probabilities. If a flop has 2x higher weight then the results on it are counted 2x. This is the same thing as probabilities. If you want them to add up to 1 then divide weight/(sum all weights).
Notice that this is the same notation as used when representing a range. For example if we have:

AA:1, KK:1, QQ:0.5 the weights don't add up to 1 either but they represent probabilities (relative ones).

Probability of reaching this line on this flop (I don't have a report in front of me so I can't give an example, if you need more information about that let me know).