Fixing a Poker Leak Part 1: Spotting and Correcting Errors

Tracking software is useful for leakspotting in cash games, but it often does not reflect some of the more nuanced influencing factors. For example, most MTT hand histories are unaware of the tournament stage. What looks like a leak might actually be a perfectly legitimate strategy on a bubble. Likewise, the leak spotter is not aware of any reads you have on an opponent that prompted you to deviate from a GTO strategy.

Coach

From personal experience, I have found that the most effective way to get my leaks spotted is to have a better player sweat my play. I occasionally record a session and have my coach, Dara O’Kearney, look at my live play. This is effective because I don’t get to cherry-pick the hands we review, and more importantly, he will spot errors that I made that were not even on my radar. If you don’t have a coach, post your hands in the GTO Wizard Discord server. It’s a great way to get feedback from studied players, and it’s a community that filters out trolls in a way that broader poker forums cannot.

My most recently spotted leak is in shallow-stacked postflop spots. If, for example, the effective stacks are 10bb and there are already 5.5bb in the pot, I tend to try and get the money in sooner rather than later by overbetting or check/raising all-in. I’m therefore giving my opponents a lot of easy decisions. Where relevant, I’ll be looking at leak fixing with this personal leak as an example throughout this article.

Trainer and Hand History Analyzer

While for me personally, getting a superior player to sweat me is the premier way to expose my leaks, there are a number of ways you can use GTO Wizard to identify areas for improvement.

There are two useful features that give you feedback where you have weaknesses quickly.

The first is the Trainer with its Drills.

This is a risk-free way to test out your gameplay and get instant feedback on your decisions. You get a GTOW score and a more detailed analysis of where you went wrong, including how to play the spot correctly. As you can see, I had a huge blunder in this hand, and the reason, on further analysis, stemmed from being too loose preflop.

A quick look at the correct range shows me that my hand was too weak to continue with as a call but was just at the bottom of my range for a bluff:

One mistake does not mean you have a structural leak, however. You can get a better idea of how you played and any trends by looking at a review of your training session afterward:

This can give you a sense of any leaks you have if the same kind of errors keep repeating. However, a more reliable way to spot them is by going into Analyze Mode and working with the Hand History Analyzer.

Here, you can upload your own hands that you have played and get every one of them GTO graded.

You can then click on the EV loss column to sort all the hands by that metric and see where you made the biggest errors.

As you can see, the most frequent spots are usually between an opener and the BB; they make up 22.93% of all spots. It stands to reason that identifying leaks in this category will have the biggest bang for your buck in the long term.

Content Database

Before putting in the reps to fix a leak, you must start by understanding why it’s a leak in the first place. If you don’t have a coach like I do, it’s worth spending a few minutes searching the vast GTO Wizard content archives to see if your leak is addressed in a video or article.

I’m not a fan of consuming strategy content just for the sake of it; I think it’s much more effective when you have a vision of where you want to improve first. So, when you have a clear leak, you are likely to pay more attention.

A quick search for my own example leak of overplaying hands in short stack spots, quickly brought me to this great free video by Matt Hunt:

Trainer/Drills

Once you have filled in any knowledge gaps, the best tool available at GTO Wizard is the Trainer with its drills. Here, you can create custom spots tailored to your own leak, so you can get a lot of reps playing the correct strategy.

This is the nearest thing in poker to a workout in a gym, and it doesn’t involve risking money.

In this example, I have picked MTT spots where the effective stacks are shallow, between 17–5bb.

Range Builder

I have a preferred way to work on these leaks, which is to use the Range Builder tool. I have written about learning in poker before; if you read that article, you might recall that I like it because it makes use of a powerful learning technique called “retrieval practice.”

With the drills, you are dealt a specific hand and given options of how to play the hand. With the range builder, you pick a spot, but then you have to input how you would play your entire range.

For example, here is BB vs UTG1, 14bb effective, the BB has checked, and the flop is 952r. UTG1 is preset to check, and now I have to input the betting hands and bet sizes.

This is much more challenging because I have to think about how my whole range wants to play. It wasn’t easy, but it is precisely how I should be thinking in these spots.

I went with something like this:

And now, I am going to write out my thought process as it went down in real time:

This wasn’t easy; it forced me to think harder than I do about a hand in-game, and now let’s see how I compare to the solution…

Ouch. I got that very wrong. How embarrassing.

Crucially, the root cause of the mistake was my assertion that this was a check-back or big bet board. This was actually a range-bet board. Looking back, I had assumed a BB call vs UTG1 open-raise was the same at 14bb as it was at, say, 40bb. In actuality, this is what it looks like:

UTG1 has a hefty range advantage, perhaps because the opening range is quite 9-x heavy, along with all the overpairs. My min-raising range is only 5% of hands; the majority of hands limp or shove preflop, so it was an incredibly tight preflop range.

Not a pleasant experience but a very valuable learning experience. As I have said before, I think the range builder may be the best-kept secret at GTO Wizard; it forces you to think about your whole range, tests your assumptions, and gives you brutally honest feedback when you are wrong.

Nodelocking

One last tool to use is the nodelocking feature of GTO Wizard AI. Most of us use this to see how to exploit opponents, but it is just as useful for showing how we exploit ourselves.

I custom-solved the hand from the range builder section with the AI and then nodelocked to see what happened when I followed that strategy.

This is the side-by-side comparison (with GTO on the left) for what happens when I bet small:

The BB raises much less often and calls a little more often against my game theory suboptimal, exploitative strategy. I think my small-bet range is too weighted towards nutted hands, so there’s less reason to check/raise.

There is no big-bet range to compare, but this is what the BB does against my big bet:

They play a very straightforward style of raising with 9-x or better and bluffing with gutshot-type hands. In hindsight, my large-bet range was also way too value-heavy, which allowed my opponent to lay down bad top pairs or worse, resulting in me missing out on a lot of value.

A really useful tool here is to compare the EV of your nodelocked strategy with the GTO strategy. I prefer to do it at the start of the hand before anyone has acted to get a more holistic view of how much the overall strategy is costing me rather than the specifics of any one node.

Conclusion

Few people can be told the correct way to play and never make the mistakes they made previously again.

Fixing a leak is about understanding why the logic is faulty and then getting the reps in, working doggedly at overriding it with sound logic.

The best way to get those reps in is by using features like the Trainer/Drills and the Range Builder. You can get very specific with your own leaks, work hard on the correction without risking money at the tables, and look at the correct solution in more detail when something surprises you.

To truly have fixed a leak, it must be fixed at the tables too. In the second part of this series, we will look at the mental game side of leak fixing, which includes how to think about executing the correct strategy in real-time at the real money tables.