Breaking Down HUD Stats with Martyn Horslen: Part 1

PokerSoftware caught up again with Martyn Horslen, who takes his HUD and poker software very seriously.  He believes that by customizing a HUD to suit his needs, he can multi-table more efficiently and increase his profit per table.  When we last spoke with him, he had his Holdem Manager 2 HUD set up line by line as follows:

Line 1 - Name, Hands, Number of Big Blinds
Line 2 - VPIP, PFR, 3bet, Fold to 3bet, Fold to 4bet
Line 3 - BB Fold to SB Steal, SB Fold to Button Steal, BB Fold to Button Steal, SB Fold vs Resteal, BB Fold vs Resteal
Line 4 - Flop, Turn, River Cbet  (in position and out of position for each)
Line 5 - Flop, Turn, River Fold to Cbet  (in position and out of position for each)
Line 6 - Fold to Flop Donk Bet, Fold to Turn Donk Bet, Fold Flop Cbet in Position to a Raise, Fold Flop Cbet Out of Position to a Raise

PokerSoftware: When we interviewed you last month, you mentioned you constantly change stats in your HUD.  Can you tell us what stats you have removed and added?

Martyn Horslen: My HUD is a huge work in progress and I am currently documenting it all on my blog.  The first stat I think is worth mentioning is that I took out the general Fold to 4bet statistic. There are a number of reasons for this, ranging from sample size and the fact that most small-stakes players are only capable of 3, 4, and 5betting with stronger or premium holdings. They would rather see a flop with their speculative hands.

I'm not sure how common knowledge it is, but I've always understood the biggest mistakes made by the majority of players fall on the turn and river, which kind of makes 4betting the least optimal line, as most players know how to handle the range of hands they 3bet with. They already know whether they are going to 5bet or fold and, therefore, the profit margins are much smaller than if you were to take someone to these later streets.

I guess the others are generally the same.  I replaced the Fold to Cbet statistics with the more specific versions, which splits the stats based on whether the hand had a single raise pre-flop or a 3bet. Pretty quickly, I saw some huge differences between the way people play in both scenarios and how it affected the general Fold to Cbet stat, especially on the flop.

I also went back over the percentages I used to color-code each stat.  I felt I had overlooked a few factors, especially leaving margins of error for scenarios where a third party still has the opportunity to enter the pot, even if the odds of that are unlikely.

PokerSoftware:  Can you tell us how you use your post-flop stats?

Martyn Horslen:  I focus heavily on flop bets because more variables and smaller sample sizes come on each street.  Everything revolves around making my opponent fold his hand.  It is usually better than mine because I'll use the stats to play any two cards when the scenario arises.  Generally, I like to be the pre-flop aggressor and so it becomes a question of whether my opponent folds to a Cbet enough.

Using similar math to the BB Fold to SB Steal scenario, making a Cbet of 60% of the pot needs to work 37.5% of the time.  I usually add 10% to this as a margin of error and to ensure some profit is made.  This makes the magic percentage 47.5%, which I will color green in my HUD.

This works in and out of position, but only against one player.  When you play against two players, you need to find the square root of this number and that will tell you the percentage needed by both players to show a profit, which is 69% each.

You can also use this for the Fold to Donk Bet statistic.  This allows you to call steals pre-flop and then bet into your opponent in the same way.

More articles