Friday, December 25, 2009

A quickie run down of Poker Tracker / PT3

Jordan over at HighOnPoker has requested a quick summary of Poker Tracker 3.0 (PT3) and its affects on my decision making during play. Since winning his hand ranging contest a few weeks back, I feel obliged to respond... hopefully it will help some of my readers, though, hopefully it will help me. You know what they [who is "they"?] say: you know when you truly know something if you can teach it to someone else. Happy belated Hanukkah to you & your family, Jordan (even though you say you're a self-hating Jew)!

Online, particularly when multi-tabling, it is hard to notice who is playing a loose style and who is playing tight. In the same vein, it is equally hard to evaluate who is playing aggressively and who is playing passively. Therefore, PT3 includes, by default, two primary pieces of information a new user should be referring to: the Voluntary Put Into Pot (VPIP) and Preflop Raise (PFR). The two gauges make up part of the default HUD setup for PT3. The two data points should be your entrance criteria for a hand against an opponent. Therefore, I have come to rely on PT3's two aforementioned stats to help assist me with my decisions. Granted, these stats are not perfect; i.e. you want to have an adequate hand sample size of 50 or more (BARE MINIMUM), and those numbers, on the surface, don't give scenario descriptions (such as raising position, tendency to steal, etc.). However, they do give a good idea as to the holdings and style in which the opponent is playing.

For example, if I see a player with a VPIP 35 / PFR 6 over an adequate hand sample size, I can safely assume that although his limping range is wide (likely including all Aces, suits, connectors, and the standard pairs, suited connectors and broadway) and his raising range is very tight (limited to pocket pairs TT+ and / or AK & AQ). When I see that 35/6 player limp, I generally want to raise him pre flop because if he calls, I generally know I'll be able to able to push him off the hand post flop (with the knowledge that his wide range misses a lot more than my more clearly defined range). When I see him raise, I'm going to be folding out everything but the very top of my range (referring back to his raising range of 6 PFR). Is this perfect? No. In live poker, I can visualize my opponent, and make reads based on what I see his physical actions are, coupled with his bet sizing and timing. However, online, we are not blessed with such tells. Therefore, we are forced to go with the hard data we have, which is the hand histories.  Hand histories are automatically saved on your hard drive when you're playing a hand.  Several sites also give you a chance to buy hand histories.

Another stat in the default HUD configuration includes a stat called aggression factor which measures a player's tendency to bet or raise vs. check or call. It is a basic measure of his or her tendencies. I put little stock in this number, however, because if a player is a 4/4 and has an AF of 20 over an adequate sample size, he's clearly playing AA, KK, QQ and should be leading / betting most flops, turns and rivers (thus making that type of player exploitable because he'll bet my sets for me, or I can dump the hand when I miss).

Besides the default HUD configuration, I am constantly looking at other data, which I sometimes include in my posts. This data can be viewed by clicking on the player's associated HUD window:
  • Cbet flop: The percentage of time the player will lead a raised flop. This number is useful in determining whether I can check / raise the flop if I'm out of position, and take note of my opponent's action when in position (i.e. if his cbet = 100% and he checks the dry flop, does that mean he hit a monster or missed entirely).
  • Attempt to Steal (ATS): The percentage of time a player will attempt to steal the blinds, pre-flop. This number is an assessment of any time a pot is folded around to the player in the CO, BTN or SB and opens the pot. At the level I play ($25NL / $50NL), this number does not seem to be all that important, as I have taken the tact of stealing north of 50% lately, without any pushback from my opponents. They either don't notice the constant steal, or don't care. Once can actually make a living (not literally) off of stealing blinds. On the flip side, I have noticed that the bigger ATS I've seen is in the neighborhood of 30%, which leads me to believe my opponents are not doing the same. However, I've seen opponents around 50%, and I know that if it's against my blinds, I can re-raise or call / lead and usually take down the pot for extra profit.
    NOTE: If you are going after blinds continually, you should note that it is tough to steal from short-stackers. Their response is a simple shove, which they can employ quite effectively. They know you're stealing, and they know you can't call a 20BB 3bet with T2o.
  • 3bet pre-flop: The percentage of time a player will 3 bet pre-flop. Critical for certain situations; I have found players who will 3bet my ATS's and I know are coming up with air, making it safer to 4bet them back. I have found players who are 3betting every raise, which tells me that they are 3betting light (and getting lots of folds). I have found players who 3bet 1-2% of the time... they're only 3betting AA, KK.
  • Folds SB / BB to steal: If I find a lot of resistance from my targeted steals, I will give up raising my BTN with 72o and range upwards, limiting my steals to a more restrictive QTo steal instead.
All of that said, I am in the process of [lazily] switching over to Holdem Manager (HEM). I have it all set up, but am experiencing a bout of technophobia because I don't want to move away from what I know and love. However, HEM can get me tendencies for each street, based on position, in a nice table, where the default PT3 layout falls short. Regardless of the tool you select, if you're playing online, you're playing at a disadvantage by not using data mining software.


  1. Why did you switch from PT3 to HEM?

    Do you think it's as valuable for MTTs and S&Gs as cash games?

  2. First question: 2 factors - I did not actually *OWN* a copy of PT3... It was a "borrowed" license (with borrowed being used liberally). Also, HEM breaks down the stats by position, whereas PT3 takes a bit more work to get there. Overall, I like HEM's drill-down displays better than PT3.

    Second question: Good question. I know that you've been reading this blog long enough to know that I'm a cash game player, first & foremost. Therefore, take my answer with a grain of salt. However, IMO there is no definitive answer. I think it probably is not *AS* useful for tourneys as for cash. My reasoning is that cash is the almost same exact scenario hand over hand with the difference being effective stack sizes.

    However, with tourneys, there are many more variables to consider how a particular player plays: the stack sizes, the bubble, table size (10, 6, HU, etc.), all of the standard variables that play into making a decision at a MTT or SNG. Therefore, a player's play should be dynamic from one hand to the next. I haven't played around with HEM, and don't doubt that it is able to do so, but PT3 can track separate stats for tourney play depending on table size. In other words, it can track a player's stats if there's only 6 players or less at the table differently than a full 10 or 9 person table.

    Ultimately, I'd imagine the answer is no - tracking software is not *AS* valuable for tourneys as for cash, but it is *CERTAINLY* a valuable tool for tourneys. There is no question. Moreover, if you multi-table, you cannot play without it.

    Email me if you want to discuss in further detail.


Blog Archive