Sorry for this (probably dumb) question:

I've came across a website: GTO range builder (http://gtorangebuilder.com/#home , mods please remove the link if I'm not allowed to post it).

They claim to build a GTO strategy with fixed bet sizes for any river spot, and if you subscribe they claim to solve also turn spots.


1) I'm approaching GTO oriented game and it seems strange to me that they're asking for ours and villain's range before solving the river.
I mean, isn't GTO supposed to work without knowing villain range and tendencies? So why are they asking for villain's range?

2) I've read in other threads that making calculations regarding a river spot is pretty much useless if the range you came to the river isn't GTO itself. I mean, if you reach the river with 100% range is pretty much useless and non-sense thinking about making a GTO decision OTR, since it will certainly be wrong. Is this true? So what's the point of the website if we cannot solve previous streets?

P.S.
I'm aware GTO with fixed bet sizes isn't real GTO.

I believe it's partly a CPU issue. It would take forever to solve if both players had 100% of hands in their ranges. The game tree would be massive if you included over 1000 possible combos for each player.

Look at the problem from the other end. If Rangeuilder had "solved" poker, you'd be able to ask it "Which hands should I open pre-flop and what is their precise EV for every possible post-flop sequence?" It can't possibly do that, because it doesn't have the computing power.

To learn an approximation of GTO, it's possibly helpful to work backwards from the river. If only a few combos can be profitably bet there, you have to work out which hands were given up on on previous streets.

Sounds like they are essentially solving a toy game, not an actual subtree of GTO NLHE. They let you customize the toy game by inputting your own ranges and whatnot.

In response to #2), I don't think it's useless. If the ranges are realistic/common, you can definitely gain some useful insight.

Hi,

I'm the creator of GTORangeBuilder so hopefully I can answer these questions for you.

The basic idea behind the software is that it is a tool that lets you do more powerful/in-depth hand/situational analysis, think of it as if CREV had a "GTO" button that would fill in a strategy tree that you built with GTO play (across every possible runout simultaneously) instead of just a max-exploit button. As a poster above said, in effect it is letting you define custom poker scenarios (extremely complex "toy games") that it will solve for you.

To see why this is useful (and to answer your question #1) think about the way you might analyze a hand or a spot where you want to improve, for example you want to improve your OOP play in raised pots on wet boards, CO vs BB.

Imagine that you have a very simple question about those types of spots that you are uncertain of or what to improve, such as "should I be leading the turn OOP on wet flops on various turn run outs". GTORangeBuilder lets you analyze these questions scientifically and precisely. In fact my entire goal with GTORangeBuilder is to try and give players to tools to scientifically study and analyze situations that people currently can only speculate about.

With existing software, the first step might be to use existing tools like CREV, odds oracle, etc you might do something like put in your pre flop range and look at hand equities, equity distributions, implied odds, etc, think about how the population in your games plays the flop and start by thinking about what kind of hand range your opponent has on the turn. This is the step that GTORangeBuilder is assuming you take when you input the starting turn ranges.

Your next step would then to start guessing at the EV of leading with various hands/ranges. The problem is that to actually determine the EV of leading for half pot with say Td9d on Qc8d2s4d is actually an enormously complex problem. First you would have to think about how you would react to a raise, how often your opponent would raise, how you would play on every possible river card when you lead and are called, how likely each river card is to come given blockers, etc. Even if you did that your analysis would be extremely incomplete because in a GTO sense, how your opponent would respond to your lead depends on the range of hands you are leading with and thus the EV of leading with Td9d depends on the composition of your leading range so you can't even really think about it in terms of the EV of a single hand in isolation and you want your leading range to result in balanced ranges when called on as many river cards as possible which is a very complex problem. Even an extremely basic question about turn play like "should I be leading when this turn card comes" is almost impossible to answer precisely with existing software.

GTORangeBuilder gives you precise answers to these question. If you tell GTORangeBuilder how you and your opponent play pre flop and on the flop (by entering hand ranges with which you might reach the turn) it will tell you exactly how often (and with which hands) you should be leading the turn to play GTO from that point forward and it will also tell you a separate strategy for how you should play every hand in your range on every possible river card. You can even manipulate the game tree by solving for GTO play where leading is disallowed and then comparing the EV of that situation to the EV of the situation where leading is allowed to determine exactly how much EV you might lose by not leading the turn. Furthermore, you can experiment with looking at solutions on lots of different turn run outs and with a variety of flop c/c ranges and flop c-betting ranges, and consider a variety of possible leading bet sizes to get a sense of how those parameters effect whether or not leading is necessary and how they shift the optimal way to balance your leading range.

By solving a turn/river scenario in isolation, GTORangeBuilder tells you "assuming me and my opponents pre flop / flop strategies put us into this turn situation how can I optimally play the rest of the hand." This is clearly a different question than "what is a GTO strat for the full game of NLHE" which is the explanation to your #1. However, given that a full GTO start for all of NLHE is not computable this is a much more practical question and is one that you can directly apply to your game to measurably and scientifically improve your play.

I have some CR vids outlining this type of process as well as some free youtube videos, you can check out an example here if you are curious: https://www.youtube.com/watch?v=iMlcwSPKcKM

With that preamble I can more concretely answer your second question

It depends what your goal as a poker player is. If you are interested in a full solution to the entire game of NLHE considering a river spot that is likely off the equilibrium path is not very useful. However, if your goal is to win more money by playing better in some particular turn/river situation that you frequently encounter in real life then it is actually going to be higher EV to restrict your study to that spot in isolation and to determine what GTO play looks like assuming that you find yourself in that situation. This is a consequence of the fact that in a full solution your strategy in an off equilibrium path subtree only needs to be strong enough to keep that subtree off the equilibrium path, which can still involve quite suboptimal play. GTO strategies are allowed to be quite weak off the equilibrium path, because against a GTO opponent, the probability of reaching that part of the game tree is 0 so those weaknesses will not allow a profitable deviation unless they are so large as to be exploitable. I have a card runners video on this topic (The Theory of Winning Parts 2/3) if you are interested. I also have a blog post that is more or less on this topic here: http://blog.gtorangebuilder.com/2014...ia-part-1.html.

I think most players would readily admit that they have no idea what actual GTO c-betting ranges and c/c ranges look like on the flop, but they also would have a good idea of how most of the regs in there games play those spots. GTORangeBuilder lets you take that knowledge about your specific games and apply it directly to strengthen your play on later streets. I should also mention that GTORB flop solutions will be launched commercial within the next few months so more information will be available about optimal c-betting (again assuming standard pre flop play) soon.

Hope this helps, ended up longer than I intended

-swc

Just to be a bit more precise on this aspect of what I wrote above...

There are two general possibilities when you analyze a specific scenario rather than the full game tree. One is that that part of the game tree is off the equilibrium path in which case everything I said above is true.

The other is that the actions taken are on the equilibrium path but the ranges with which actions on earlier streets are taken are not GTO. The result in this case is very similar. By solving a model scenario of those later streets in isolation with the actual ranges that people use to reach that spot in your games, you will end up with a "tilted" equilibrium. This means that its possible that someone could adapt their play on earlier streets and increase their EV, but that if they play earlier streets as you assumed then there are no possible +EV deviations. Furthermore, your EV in the situation where people play early streets as you assumed will in general be higher than your EV where you play a GTO strat in the full game of NLHE as the GTO solution to the later streets in isolation will increase its EV by taking advantage of the suboptimal way that people actually are playing on the earlier streets, which is a concept called "tilting".

-swc