You can only play versus a specific decision if "Play versus selected tree" is selected. If playing versus a database, it will be a different decision for each tree.

It makes perfect sense to me to let users build their own trees. However, it only seems reasonable to let people know that, as far as I can tell, in GTO all bet sizes perform similarly. So to let people know that this is probably not a very productive area to be focussing their time and efforts on.

Ok, I will consider it for later releases.

Sometimes it helps to see how bad a certain choice is. For example, if the correct action has an EV of 11.10 and the incorrect one has an EV of 10.90 then choosing the wrong action is a small mistake. But sometimes the EV of the incorrect action is far less than the correct one, making it a large mistake. We display the EV in order to give this feedback to the users.

EV is always expressed in $ (or €,£,etc). BB/100 would require us asking the user what the BB actually is, which would make the interface more difficult to use. In the end, EV is not an overly important part of a GTO solver. It mostly serves to show how good/bad certain decisions are. It's more a qualitive measure of play than a quantitive one (meaning that the exact values or the units are not that important; what's important is which is the bigger number, and by how much).


Yes, we haven't added that yet, but it will be available in later releases. With GTO+ v120 and FlopzillaPro both having been largely finished we decided to first release a beta and take care of the final details at a later stage. The alternative would have been to delay the releases by a few months.

This feature is intended for storing predef ranges. To use it, turn ON "Apply groups to selection". Then, select any groups that you like in the matrix. When you're done, click on "Add range" to store your grouped range to the predefs.

The reason why I'd prefer a running tally of EV performance is that it isn't subject to variance, whereas a dollar amount isn't really a good measure of real performance, especially over a small sample.

Basic question on GTO+

The introductory vids assume a lot more familiarity with this sort of software than I have.

I understand the program offers unexploitable strategies and the tree depicts decision points throughout a hand. But I am having difficulty picking out relevant details. The demo sim I ran seemed to offer many choices, all to be done at 50% frequency, which seemed odd.

I'd like to see a very basic vid -- this right there is the info saying to do what.

And we're just telling you that nobody else seems to hold this opinion. Not including bet sizing recommendations and then telling us it's because they're "not relevant" just doesn't really cut it lol. I love the product but I look forward to seeing you implement what you now know your customer base does consider relevant and your #1 direct competitor already has in it's product.

There's zero reason for it to not be in there other than just cutting corners.

I don't understand what I can do here anymore.
I have explained my motivations in my previous posts.
Other people have tried to explain it to you as well.
For example see posts #8498 and #8500.

If you want to continue this discussion, can you provide me with an example of where you feel bet sizing affects the overall EV in a significant way? For this, just send a savefile to support and I'll take a look at it. Please realize that all of this focus on bet sizing is pointless if you miss just one value bet, or make a single incorrect bluff. The value of bet sizing is so tiny compared to the value of quality of play that I feel it would be dishonest to our users to add a feature that would steer them in a direction where there's little to be learned.

Ok, I will take it into consideration for later releases.

Can you please watch the last video here: www.gtoplus.com/videos
It starts with a very basic description for FlopzillaPro users.

I understand and agree with the overall EV of betting scylla. That's one of the fundamental concepts of mixed strategy equilibrium. But you said that Pio added betting suggestions because people asked for that and not because it is useful . How can you say that, if we don't actually know how pio is making these suggestions(theory wise). If pio is randomly suggesting bets then yeah it is not useful. But if there is some theory and formulas behind that then i guess we have to at least study it to understand if it is useful or not. (and by the way it is a product for sale. It obviously needs to listen to people demands lul)

All of the above being said, this feature is basically already available. For that, just create a database of the same tree, but for different bet sizes. And then solve the database. I did this before in post #8490. See pic below. The first column contains the % of the bet size. Come to think of it, I think I actually demonstrated this in the second video here: www.gtoplus.com/videos



And if you want to run the same database, but for a different flop, then just use the option "Change all flops to".

In post #8487 I explained that bet sizing has almost no influence over EV performance when playing GTO.

In post #8488 the question was asked "Why does pio suggest bet sizings then?".

In post #8490 I explained that this is because their users requested it. Similarly in post #8511 above you stated about GTO+: "and by the way it is a product for sale. It obviously needs to listen to people demands lul". Which was the entire point I was making in post #8490. Namely that often features are added because users request them; it does not necessarily imply that they are useful. The entire point of this discussion is that my statement that bet sizing does not seems to affect overall EV performance is not disproved by pio having a feature based on it.

That being said, as explained above in post #8512, if you want to research bet sizing, then you can just make a database for the same tree, but with different bet sizes. See video 2 here: www.gtoplus.com/videos. So, we ourselves have already added functionality for this purpose. However, if you want to use it, then please observe that I don't see any reason to expect much to be found here. The influence of bet sizing seems to be typically around 1% of the overall EV. A single wrong bluff or missed value bet will negate any gains made with bet sizing. When playing GTO, the importance of quality of play is far greater than the importance of bet sizing.

I have a technical question. I use a pretty cheap laptop for poker. The RAM is pretty low and it can not be upgraded. It supports a HUD and poker room but not much else (at the same time).

I understand GTO+ requires a good amount of RAM to solve for spots but does the CardrunnersEV part of the program require a lot of RAM to use?

If you're referring to CREV's GTO solver, then it will require similar RAM to GTO+.
A bit more in fact.
Without the GTO functionality I'd say that you need about 1.5GB for CREV at most.

I'm not aware of any feature where pio does "bet sizing recommendations" beyond the equivalent of the above (running separate sims of different bet sizes and comparing the results). So I'm really confused what people are talking about.

All I can really say is that, when it comes to ourselves at GTO+, we have designed our interface in such a way that it's possible to create all sorts of databases. This includes the ability to create a database with trees of different bet sizes. So if our users want to look into bet sizing, then they can do so.

I wasn't aware of another product offering a dedicated feature for this, but the two posters above seemed rather confident of it existing. Admittedly, I haven't bothered to check, and I can only blame myself for that. All I stated that the existence of such a feature would not disprove my statement that bet sizing has only very little influence on long term winrates when playing GTO, and that I would recommend not investing too much time looking into it. Any reasonable bet size will perform similarly to any other reasonable bet size.

I suppose a point of improvement for the future would be that, even if mulitple users ensure me that some feature exists in some other product, I should still check to see if this is actually true, instead of taking it at face value.

Pio doesn't have any feature of recommending bet size and I'm not sure it would even be useful. You need some framework to understand why and when a bet size is optimal or accomplish something regardless as memorizing bet sizes and learning how to execute strategy by rote is really inefficient and not a strong human approach of learning.

I very much agree with Scylla's view that picking correct turn/river bet size doesn't matter all that much. Linus pretty much used exactly that approach on his way to the top of simplifying bet sizing but learning to play it really well. This other coach mentioned isn't really much of an argument for much as he's just an average 100z or bellow average 200z reg. There are some spots having a big size is important and some very specific spots a small size is important but usually just a mid size like 75% pot for bets and raises practically gets all the EV.

OK so here is a simple scenario: you are up against someone who just loves to play low suited connectors and any two suited cards. You get a heads up versus him with Ace Ace and the board flops 568 two tone. Your strategy is to always bet 33%. He is going to call with any draw because he knows that when he hits his draw he’s going to get paid off when he bets 200% pot because you always call because GTO says villain has enough bluffs in his range to do so. It seems like you would get wrecked this way.

Seems like a big GTO mistake to cbet entire range with 33% pot on that flop (would guess correct frequency is more like 25%) if villains range is a lot of mid-low pairs and suited connectors. AA should always check against a decent player too so would never get into that spot if you played like solver.

No one is stopping you from inputting that 200% size either.

I'm not affiliated with GTO+ but I saw a lot of misguided stuff here today and I'm subbed to the thread after making a post a week ago so just popped in with a couple of posts. Scylla seems to have a good vision of what makes a useful solver at this point and the product is headed in a nice direction and is very useful already.

I created a scenario where a TAG (OOP) opens with TT+, AQ+ and a LAG (IP) calls with a super-wide range of [JJ-22,AJs-A2s,KQs-K2s,QJs-Q5s,JTs-J5s,T9s-T7s,98s-96s,87s-85s,76s-74s,65s-63s,54s-52s,43s-42s,32s,AQo-A2o,KQo-KTo,QJo-QTo,JTo,T9o,98o,87o].

OOP betting size is 33%. IP betting size is 150%.

Starting stacks are $1,000. If IP bets each street, stacks are in by the river.

On a flop of 8s 6d 5d, OOP is checking 100% of range.

After OOP checks flop and IP bets, OOP folds 60% of his range including most of his TT, JJ, QQ, and even some of his KK!

If OOP calls a flop bet, and the turn is a brick like 2h, he donk bets 100% of his range.

If LAG 3-bets the turn brick, then TAG 4-bets and basically is willing to get stacks in with an over pair and a few nut flush draws.

If the turn is a little more coordinated like a 4h, then OOP checks 100% of range again. IP now bets nearly 100% of range and OOP folds further 62% of his range including most of his QQ and KK. He's left with a few AA, KK, QQ and nut flush draws (calling two 150% pot-size bets with flush draws, really?).

If the river is now slightly coordinated again like a Th, OOP checks 100% and if IP bets, he folds another 60% of his range.

It seems like there's a lot to figure out from this scenario.

My first question is, is there any significant flaw in how I set up the scenario parameters?

H Scylla. A +1 to this request please. I use descriptive filenames but it is often insufficient.

Great job.

Best regards
Rod

What is CREV most useful for now that GTO+ and FZpro are available?

GTO+ is to a large extent CREV2 and FlopzillaPro is Flopzilla v2. That being said, CREV is still useful for learning EV calculations when studying, visualizing ranges, as well as offering several other tools for calculations. Purchasing CREV includes GTO+ though, and vice versa, so purchasing either product will also get you the other.

Ok, I will see what I can do for later releases.