Are there keyboard shortcuts for navigating the viewer?

There are shortcuts for some functionality. You can see description of them after going to the options in the top menu. The shortcuts are next to the function, like that:

hey guys, I'm a complete noob with computers and currently looking to buy a new laptop. I'm wondering what comes into play when making pio calculations. Is it just the processor and ram? Which laptop would be best for me to make the fastest pio calculations, I'm currently using an asus with an i7 and 8gb ram but when I do calculations with wide ranges (bvb for example) it takes awhile for it to get the results.

So yeah which specs come into play here and any recommendations for which laptops I should consider?

CPU to make calculations faster. RAM to be able to build bigger trees.
Sadly latops are not ideal for heavy computations but if that's your only option you want:

1)a real quad core CPU (that rules out ultrabooks with U or M series CPUs, you want a real quad so HQ series)

2)You want 16GB of RAM to future proof it. You won't be able to run a preflop solver on that. 8GB might be enough for the postflop one but you will run into limitation if not today then in the near future and not only with Pio. Don't buy a computer with only 8GB of RAM today - period.

You are very limited when it comes to CPU choices on laptops. If you already have a quad i7 (double check, some are dual core) you can't get much faster. Maybe it's better to wait till we have software improvements instead

thanks for the info, are there any laptops you would recommend in particular at a $5k> price range? I'm thinking about a desktop as well but since I travel a lot it may be a bit of a hassle, which desktops would you recommend though?

I don't really know much about the current hardware market I specifically don't know anything about laptops (other than that there is no ultrabook with a quad core CPU).

I would consider renting a dedicated server. You can find something <100$/month which is going to be better than a decent desktop (say hexa core CPU with 64GB of RAM) or you can spend more and get a really big machine. It's very convenient to use because:

1)you can log in to it from any place and any computer
2)it doesn't use your local resources during calculation
3)it's a convenient storage/computing place for other things than a solver as well

While there is a bit of a learning curve when setting those they are dead easy to use once you go through it.

I've used an Asus G751 laptop with i7 4700hq processor for my calcs. It's a gaming laptop so if you want portability and a good CPU those are the ones to go for. You can get a 7700hq or 7820hk processor nowadays on most of these. Asus, MSI, or if you want more of a "business" look go for Sager as they look less game like. They also tend to cost less and some even have desktop processors in them like the 7700k (https://www.sagernotebook.com/Intel-Core-i7-7700K/). Before purchasing just make sure to check notebookreview forums and notebookcheck on the model you want to see how it performs temperature wise and try to get them to apply thermal paste to the CPU that isn't factory paste. I've ran multiple 1755 sets on the above Asus that is now a couple years old and have had no temp issues but I did get a proper CPU/GPU paste in the beginning. It does a Pio bench of ~9

I might as well report following my upgrade.

I went from a lightly overclocked quad i7 6700k @ 4.5 GHz to the octa core Ryzen 1700 @ 3.7 GHz the other day with PIO in mind. Both running equally cool with similair power draw at 1.250 volts. 6700k yielded 6.8 seconds in the PIO bench, Ryzen 4.3 seconds. You can push it into the 3.8-3.9 second territory but I prefer low voltages for a cool, quiet and effecient setup rather than maximum speed.

At $300 it's fantastic, with the caveat that you'll need to do the overclock yourself. If you want something plug and play the 1700x/1800x is still pretty amazing value for money.

Hey there!

Could someone please give me top 2-3 use cases for the preflop section?

Thanks

Yeah, Ryzen is very nice when it comes to Pio performance and definitely offers the best price/performance on the market today.
Some of our users experienced problems with new Ryzen setup and it seems it was cause by memory instability. My advice is to choose your computer store carefully so they are able to choose good components and set everything up properly.

In basic/pro version that's basically a toy which can be used for:
-solving pure preflop trees (assuming checkdown postflop)
-constructing preflop configs and estimating their size (before you buy edge version to see how much RAM you need to solve them)
-load preflop saves from someone who run them with the preflop solver

In the edge versino you can solve preflop spots which are already HU, so:

-SB vs BB
-normal HU
-BTN steal with defined range vs BB

etc.

Yes, AMD is just now rolling out better memory support to the motherboard vendors. I went for a higher end motherboard precisely because it was one of the first to implement this.

There are also reports of instability when reaching about 75c. I can concur since I had one or two failed stress tests with the original cooler, but since I got the mounting bracket for my Noctua D15 the same settings have passed everything I've thrown at it, including all the common stress tests and a few multiple hour PIO runs.

ran an aggregated report for my 184 flop subset, the last line is what im assuming the average of the results? but when i calc it myself i get a different average. what do these #s represent?

I have a question, when i choose multiple bet sizes (on the flop for example) and the program really favors one betting size, lets say A97 flop, if the program favors pot size bet with 100% of AK AQ and AJ.
My questions is, when i choose to continue the hand analysis choosing 50% bet size, on the turn i will not have any combo of AK AQ and AJ because i lost them not choosing 100% pot.
How do i manage to control this? The only way is to rebuild the tree with only the most suitable flop bet size?

Those are weighted averages. The problem is that you need to include both weights of flops and matchups (probability of things happening) in the average so it's not that easy to calculate by hand.
As simple example why matchups are important, let's say you run a report on two flops:

A73 and A72 and you look for decision for 33.
Let's say the solver always bets with a set and always checks with an underpair. What should be an average for a bet?

The answer is 33% because you can have 33 in 6ways on A72 but only in 3 ways on A73. Those probabilities include also opponent's range btw so those things are not easy to calculate by hand. The report should include matchups though so you can verify that. My advice is to run very simple cases (like 2-3 flops) for manual testing as it's too easy to make mistake on big subsets if you are going to calculate things manually.

Yes but the way the solver works is that the strategies for AK/AQ/AJ are still at least decent in those branches. You can show strategies for hands not in range by unchecking "square sizes proportional to weight" checkbox, here:

https://gyazo.com/15de0348aec88f37291de8a3ed887e20

Hi, i got some problems with piosolver. I built a tree and is runs until its solved. But i doesnt give me no more turn and river cards. Is says "final node"

Well, final node means that the tree ends there. The reason might be that both players are already all-in or that a branch following that got removed.
To say more I will need:

-the tree config (big "copy to clipboard" button in the treebuilding calculation tab, paste to pastebin.com and link)
-screenshot (or description of a line) from the place in the tree when you expect it to go further but it ends in a final node

Hi guys, I have the Pro version of PIO and i ran into a weird problem.

When i gave it a simple task, like solve for IP range of: 99,66,Q9,87 and OOP range of 93o

on Qs9h6d it is reporting in decision tree that IP has 31 combos to make a decision with over all streets instead of taking the blocker effects of 93 and reducing it to 25.

Is this standard or am i somehow bugged? It is not able to discount the blocker for value range nor draw range(if i take like 98 for OOP hand and it reduces the combos of 87)

thank you

I am assuming you are talking about this:

https://gyazo.com/2b7cb102cf3e22ae8480ae6e07862850

It doesn't take into account card removal from opponent's range. This is by design. It's a better way to show things because otherwise it makes little sense to speak about "combos". For example if the opponent's range is typical stealing range we would need to adjust 6 combos of AA to something smaller (as the opponent's range is A heavy) and 33 to something bigger. You often don't know what range the opponent exactly have so this would be very confusing.

There is an option to show overall frequencies for actions, those ones:
https://gyazo.com/dbbc4bdc50574c6344d5cc08f6e71425

as real frequencies (how often it really happens) by going to Tools->Configuration->Data presentation on 13x13 grid->show total strategy as real frequency.

I talk about the difference and why the default was chosen the way it was chosen in our video here:
https://www.youtube.com/watch?v=zLGpcZavxeQ (16:10)

In general there are two ways to show things:

1)what the player thinks they are doing assuming they don't know opponent's range
2)what is really happening if both players follow their strategy

1)is in general more helpful because you often can't know opponent's range exactly (at any place in the tree other than the very beginning).

2)would become very confusing very fast in situations where the opponent's range is a bit more wide than 93o.

There are two files in each subfolder of preflop_subsets.

Which is the difference between this two files?




Thanks

Those are just 2 different subsets of 487 flops. We've recently run another set of benchmarks and our view is that the best subsets are the ones listed here:

https://pastebin.com/AtyU84qC

Is it possible to run multiple scripts at the same time? If so, if I run two simple scripts of equal size at the same time, roughly how long will it take compared to running one after the other, note: I have a really powerful computer.

Yes. The viewer opens another solver process to run a script so there is nothing stopping you to run more than 1.

It should be about as efficient. Maybe you can get it a bit faster if you run two but that won't be true if you run really a lot of them. Please remember that 2 scripts at the same time will 2x amount of RAM so you need to have that available.

Few quick questions on dedicated servers, hoping someone can help me:

1. What specs are generally going to be required to run a server with PIO on it?
2. Will connecting with wifi work fine?(I assume so, just being safe)
3. Can I use a MacBook to connect to a server running windows with PIO?
4. Is there any downside to using a dedicated server?


Thanks

Hey, I have a problem with the aggregation report. When I run it I receive all the data in one excel column, so it is impossible to read it. I attached a picture about it:
https://gyazo.com/6a33cc2b714c5ef8a9ad392ec0b6105e

What should be the solution for that?

All you need is modern 64bit Windows.

The standard way to use a server is to connect using remote desktop interface (built in in Windows). The way it works is that you just have a window with a server "in it". It's dead easy to use after the initial hurdle of setting things up.

I've never used a Mac so I am not 100% sure but I suspect using a Windows server from a Mac is possible.

1)You pay per month but you end up not owning anything
2)You need internet connection to use it
3)Transferring very big files (especially uploading them) might be a bit problematic depending on your internet speed/data plan

In general I recommend renting as you can run the solver 24h/day, it doesn't clog your resources and you can connect to it from any place and any computer so there is no need to build powerful hardware at home.

This is a bug in Excel. I've explained it here:
http://forumserver.twoplustwo.com/sh...postcount=1666