Opponent Types: Short-handed Stats - Recommendations for Small Stakes Hold'em

Introduction

In this article

• The most important stats for short-handed
  
• Recommendations for your own game
  
• Stats have to be interpreted
  

Tracking programs like PokerTracker or the PokerStrategy Elephant record all the important values of the sessions. This way you keep tabs on your winnings and losses, and on the other hand you have the necessary database to scrutinise your own playing style empirically. It enables you to find leaks by comparing the values with those of other winning players. Do you play too loose? Too tight? Are you aggressive enough? Or even overly aggressive?

This article is meant as a guide for the different statistics. The second part will amplify on how to include the stats on the poker table, which means what role the stats have when determining your opponent's style.

There are no optimal stats. Poker is a situational game. Not everybody is facing the same opponents. They vary from platform to platform, limit to limit, or even at different hours. Different opponents require different strategies to exploit their weaknesses.

Furthermore, not everybody plays equally well. If you know that you don't play very well OOP and without initiative, it might be more sensible for you to defend your big blind a bit tighter. If on the other hand, you can handle marginal situations well enough and know where the weaknesses of your opponents are, you can play much more liberally.

An important value with empirical data is always the sample size. Some values like your own VPIP or PFR are quite accurate as you often have the opportunity to invest money into the pot, and also often take this chance. After 10k hands, these values are already quite trustworthy.

In contrast, special values like e.g. Folded To River Bet need a larger sample size as you are significantly less often confronted with a river bet.

Thus, if you evaluate stats, you should always keep in mind the sample size. Some values are already quite sound in a relatively short time, others need quite some time to converge. The best example is unfortunately the win rate. More about this later though.

The own Stats

VPIP is an acronym for “Voluntarily Put Money Into The Pot”. This means it's an indication expressed as a percentage of how often you are willing to voluntarily invest money into a pot. Every call and raise is counted. If you get a free flop in the BB, it's not counted, as you didn't pay the blind voluntarily.

The Pre-flop Raise (PFR) tells you how often you raised pre-flop obviously. If you have a PFR of 20 after 100 hands, then you raised 20 times or, if there already was a bet in front of you, you made a 3-bet respectively. If you raise pre-flop, you voluntarily put money into the pot and thus the PFR can never be higher than the VPIP.

Example: A player with 25/20 stats invests, on average, in 25 times money into the pot and raised 20 times overall in 100 hands. This means that in the five cases in which he didn't raise, he either limped, cold-called or defended his BB.

According to the PokerStrategy charts, VPIP 24 and PFR 18 are good standard values. This is only the case though if you mainly play 5- or 6-handed. If you also play ultra-short with 4 or less players on the table, you receive higher values. You should then filter to 5-/6-handed to get a better comparison. The same applies to almost every other value as well, like VPIP SB, Folded SB/BB, WTS, ... .

VPIP/PFR are also dependent on the opponents. On a table with many loose and passive opponents you could overlimp a bit more often, whereas on a table with very aggressive opponents you often have to play 3-bet or fold.

If the VPIP drops below 20, you play too tight; a VPIP above 28 is too loose. The PFR should always increase accordingly. 22/16 - 26/20 is still within the limits. You should avoid stats like 26/16 or 22/20 though as you then either call too often or too rarely.

The values for ultra-short games are only mentioned as a rough guideline. SH beginners should rather not play ultra-short as you are even more often in marginal situations in this variant. Furthermore, “the” optimal stats diminish more and more the fewer opponents sit on the table, as you adapt more and more to the remaining ones.

4-handed: VPIP 35 +-5, PFR 26 +-3
3-handed: VPIP 50 +-8, PFR 38 +-5

The VPIP SB depends a lot on the opponents. If they are very loose and there are often limped pots, then you get good odds and can complete many hands. Even against only one bad playing limper you can still complete a marginal hand, especially when the big blind is passive. This way, you can get a VPIP SB of up to 40 in very loose games.

In aggressive games in which there are often raised pots, you receive far worse odds. As you have to play OOP post-flop, you should only play those hands which are strong enough for a 3-bet. In such games against aggressive or decent opponents respectively, you can expect a VPIP SB of about 30.

In average small-stakes games, the VPIP SB is around 33.

Naturally, this ratio changes depending on the blind structure. On 5/10 with blinds of $2 and $5, you should play a bit tighter. It isn't very easy to find empirical reference values as the opponents on 5/10 are on average also better than on 2/4, and thus the VPIP SB will automatically decrease anyway. In general, it will be about 3% smaller than on the 0.5/1-2/4 limits. If you had a VPIP of 33 there, then it will be around 30 on 5/10.

On 3/6 with the structure $1 and $3, you should play even tighter. You not only complete tighter, but also open-raise less often from the SB. The VPIP SB is about 5-6% below the one of the normal blind structure.

Against an open-raise from CO or BU, you should fold your SB in 85% +-5% of the cases, albeit a value close to 90 would be rather too tight. It also depends on the circumstances though. When you know that your opponents raise rather tight and don't know much about position, you should defend the SB somewhat tighter as their raises are more often legit.

If a LAG open-raises you can play much more liberally. If you are well versed in post-flop games you can push this value down to 80. You should keep in mind though that it could at times be tough to play OOP post-flop against a LAG, as he knows how to use his position. If you are insecure in such situations, you should rather 3-bet tighter.

When defending the SB, you almost always play 3-bet or fold. Due to the worst-possible position, you want to secure the initiative post-flop and at the same time press the big blind out of the pot. If we only called, the big blind would receive odds of 5:1 and have position on us. Thus he could call many hands +EV, which costs us value. With a changed blind structure the Folded SB To Steal changes as well.

The big blind defence depends on a series of factors. The most important ones are how you estimate the opponent's range, how well the opponent plays post-flop and how well you play post-flop yourself.

If you aren't trained in post-flop play and rather want to avoid marginal situations (especially when the opponents steal rather tightly), it is OK to start with a value of 65%. You should try to push it down to 60% as quickly as possible though. With more experience in the blind battle, you can even drive the Folded BB value down to 55.

Against loose stealers, a Folded BB value of 50 is desirable. Very good players can defend the big blind very liberally in fact. On the mid limits where the rake is rather small and the opponents very aggressive, a Folded BB value of 40 and even lower is possible. For inexperienced players, this isn't advisable though: 55% is a good point of reference.

If the hand becomes multi-way because there is another cold caller between the raiser and us, then we have to defend differently of course. For pot odds of 5:1 you can play many multi-way hands. Any suited hand, except perhaps trash like 72s, and many offsuited connectors. Hands with reverse implied odds and bad playability like K6o should rather be folded.

The gap between VPIP and PFR mostly comes from the BB defence, as we barely limp but mostly play raise or fold. Should the gap between VPIP and PFR be unusually big or small, you should check if the Folded BB To Steal might be too big or small as well, and adapt your Defending Standards accordingly.

An Attempt To Steal is any open-raise from CO, BU or SB. From experience, the ATS is approximately the BU range of the stealer. According to the ORC, you open-raise 37.7% from BU. The ATS won't be that high though. According to the PokerStrategy Charts you will get an ATS of about 35. If you want to loosen up your stealing range a bit, you can also increase this value to up to 40.

Poker is a battle for the blinds. Without blinds, there'd be no reason to even raise. It is the dead money which also makes hands which are worse than AA profitable. The more dead money there is, the more hands become profitable.

In return this means that you should steal tighter at changed blind structures on 3/6 and 5/10. You should mainly scratch the marginal hands on the lower end of the range, especially trouble hands like K7o or also A2o from the button.

In CO you can also scratch JTo on 3/6. These hands are already only marginally profitable on normal blind structures and lose a lot of value when the pot becomes multi-way. Nevertheless, the ATS should not be more than 3-4% below the one at normal blind structures.

By tendency, marginal hands have more value the fewer players there are in the hand. Therefore, you will have a higher WTS value in short-handed play than in full ring play. A value of 33% is probably a sensible minimum, 42% is about the top limit; but you have to keep in mind that you only get over 40% in aggressive games. A WTS of 33% is usually too low, except you have a certain opponent situation like e.g. very predictable opponents who are only aggressive with good hands.

A recommended WTS value for small stakes is 35-38.

On the more aggressive mid limits, a WTS of 38+ is better.

For a WTS of 35-38, a WSD of 51-55 is realistic. This value is mainly dependent on the WTS. With a high WTS, the WSD decreases, and with a low one, it increases. If you make many tight folds on the river, and also in general rather fold than call down, then your WSD will increase.

Often you do receive good odds for a call down. If you have a very high WSD, you probably fold too often. You should also include the values for Went To Showdown and Folded To River Bet. Of course the opposite is also true. When your WSD is very small, you probably make too many loose calls.

The WSD is also influenced by the sample size, similar to the win rate. A player in a downswing most often has a rather low WSD and a player in an upswing could generate quite high WSD's. With a small sample size (20k hands) you should not rely too much on the WSD.

The pre-flop aggression is mainly reflected by the PFR. Therefore, the PF AF is rather useless and should be unchecked, as it will otherwise be used in the calculation of the overall aggression. In PokerTracker, go to the “General Info” tab and there to “More Detail...” and remove the check mark at “Include Pre-flop numbers in Total Aggression Factor calculation”.

As the TAG style does not only consist of tight, but also aggressive play, it means that our AFs are probably high. The AFs are calculated from (Bets+Raises)/Calls. As we are rather tight, we'll call less often, and as we're aggressive, we'll bet/raise a lot.

The AF is highest on the flop and decreases towards the river. On the river for example, the pot is already very big and we will only rarely make a tight fold, and often make a crying call.

Our bets and raises will also become thriftier as we only bet for value or bluff. The protection concept is cancelled on the river. Good AFs for flop/turn/river are 2.8/2.3/1.8. On small stakes, somewhat higher AFs like 3.3/2.6/1.9 are also good as we can often make good folds on flop and turn (

→ less calls), and our opponents are rather passive, so that most of the time we take over the betting/raising.

On higher limits, the opponents become more aggressive and it is often more profitable to give up the initiative to induce bluffs. Therefore, the flop and turn AF can also be a bit lower. A glance at the WTS also helps. If you have a rather high WTS, it's ok if the AFs are a bit lower, whereas you expect higher AFs with a lower WTS. On the other hand, a high AF can also cause a low WTS because when you play aggressively you'll finish more hands without a showdown.

Don't forget to bet for value on the river. In the past it was said that a river AF of 2 would be good, but this is hardly possible. On the river you're often forced to make crying calls with which the AF decreases automatically, even if you make thin value bets.

An AF of 1.8 is very good. If the River AF is only 1.6, you either don't value bet often enough or let profitable bluff situations slip respectively. On the small stakes, the often missed value bets are the problem rather than the missed bluffs.

The Total Post-flop AF in a loose-passive game should be between 2.3 and 2.8, and in aggressive LAG games between 2.0 and 2.3.

Nevertheless, the AF stats are the ones you should handle most carefully. Don't change your aggression only to make your stats look better! Rather take a look in the sample hands forums and learn in which spots aggression is suited and in which passivity.

40 and less is good. If you are way above 40, with an according sample size, you probably fold too often. The pots on the river are big and even passive calling stations tend to bluff donk the river at times. As they often go very far with very marginal hands, they also more often than not stand at the river with busted draws.

A TAG would have folded his gutshot draw or overcard on the turn already, a loose opponent is rather unimproved on the river and tries a desperation shot at the win with a bluff. As we get good pot odds and don't have to catch a bluff very often to be +EV, we should not fold very often on the river.

A tighter style pre- and post-flop makes sure that we have above average hands on the river. You can only take the liberty of tight folding regularly when you have very good reads. Otherwise, you have to simply pay off the opponents. Furthermore, many opponents on the small stakes miss out on thin value bets, which is why their betting range is polarised towards strong made hands or bluffs.  This in turn makes it a lot easier to spot and catch bluffs.

Of course you can also exaggerate the calling and make too many hopeless calls. This can be found in a too low Folded To River Bet, a too high WTS and a low WSD.

Let's be honest, this is what it's all about. You ask yourself all the time: How good am I really? The disillusionment beforehand: The win rate won't help much.

Why not? Because you need a very high sample size for it to be somewhat accurate. A typical standard deviation on 6max for a TAG is 1.7 BB/Hand. For 300 * 300 = 90k hands, the standard deviation lies at 300 * 1.7 BB = 510 BB.

This means that a sample size of 90k hands is within a 70% confidence interval of +- 510 BB of the expected value. Within 30% it might even be outside of this interval! And 510 BB on 90k hands correspond to 0.56 BB/100.

Thus you have to play at least 100k hands to be sure about it up to 0.5 BB/100. Even then you cannot be entirely sure. If you want a 95% guarantee, the deviation even lies at 1 BB/100. 

The alternative: The win rate only helps partly with the determination of the playing ability. What does help then? Comparing the skill of your opponents with your own!

If you see that your opponents often make costly mistakes which you avoid yourself, then you will also play +EV. The worse opponents you find, the better → Good table selection.

Furthermore, it is important to improve your own skills. This way, you increase the skill gap with the competition or retain the skill lead resp. when the opponents become better (e.g. after going up a limit). The best methods to improve the playing ability and to find leaks are: Sample hands evaluations, coachings, videos, discussions with users, etc.

Concrete values: Everything that is positive is good. The majority of the poker players is made up of losers. Even as a break-even player you have managed to beat the rake, which can already make out up to 3.5 BB/100. A win rate of >1 BB/100 is good and a win rate of >2 BB/100 is very good.

Of course it depends a lot on the playing ability of the opponents and the rake. In especially soft games it's also possible to have win rates of >3 BB/100, although you hardly find those anymore (nowadays). A player with a win rate of 1.5 BB/100 could already be considered as a strong player. But as mentioned before: Keep an eye on the sample size!

Opponent Stats

A very important point for the determination of the opponents' stats is the significance. At times you only have a very small sample size of maybe 50 hands for your opponents, and you ask yourself how much you can trust certain values. 

Example: After 100 hands the observed VPIP differs about less than 10 from the actual one in 95% of the cases. Thus you can already say that somebody with a VPIP of 60 is already a big fish (50<70), whereas a TAG with VPIP 25 could also be a rock (VPIP=15) or a semi-loose fish (VPIP=35).

Warning: If you like to have exotic values like "check/raise turn" or "won SD when raised turn" displayed, you should keep in mind that the sample size has to be quite substantial to receive sound values.

For example in the author's PokerTracker, he only raised the turn in 1.5k of the cases in the last 80k hands. Thus the situation occurred only once every 50 hands. Even then, it also had to come to a showdown to change the counter at "won SD when raised turn". For a sample size of only 50, you'd already need over 3k hands!

Hint: To quickly get a bigger database, it's worth taking a look at the Databases Forum.

Somebody raises from CO and you want to know which hands you can 3-bet from the BU. Here the range of the raiser is crucial, but stats don't give us any actual statements about the range. First of all, we only know the PFR, which is composed of different positions and situations though: Many players raise looser on the BU than from MP2; one person raises tighter against a limper than normal, another person looser by contrast.

Second, nobody tells us which hands are being raised. The PFR only tells us how often a raise took place, but it doesn't tell us with which hands the opponent raised: One opponent might raise the Top 20% from the Equilator, another one always the Top 10% and the 10-30% only in half of the cases.

A calling station with 4 PFR does not necessarily raise the Top 4%. Often, fishes with a low PFR raise when they feel like it, regardless of their actual hands.

To imply the range based on the stats, another transformation process is still necessary in which special assumptions about the opponent slip in. If an opponent with a PFR of 18 and an ATS of 35 open-raises from the BU, it is sensible to assume that he plays according to the ORC range, but you cannot know it for sure.

It is merely an assumption, in which we transform the stats into a range with the help of logic, experience, history, etc. And whoever assumes the best in poker, also wins the most money.

If you know what a certain action of an opponent means, then you have a read on him. Stats are a part of the reads, but definitely not the only ones. If you know that your opponent overplays marginal hands like bottom pair, it's more useful than his AF. To know if an opponent caps his draws on the flop for free cards, if a passive fish semi-bluffs with a strong draw, with which hands our opponent donks into us; that are all reads. Such information cannot be won via stats, but only by observing the opponent.

The stats VPIP, AF and WTS are not independent of each other. An opponent with a high VPIP has on average weaker hands on the flop than an opponent with a low VPIP. When both players play identically loose or tight resp. post-flop, then the player with the high VPIP would have a lower WTS.

If for example a 90 VPIP fish and a 10 VPIP rock would play fit or fold on the flop, then the fish would have to check/fold a much bigger part of his range than the rock. The reason why many calling stations still have such a high WTS is because they also play much too loose post-flop, and because they play too passively.

→ if a player has a high VPIP and a high WTS, he plays extremely loose post-flop.

A high AF leads to a low WTS. First of all, you win more pots without a showdown due to the aggressive game and second, the AF is calculated by (Bets+Raises)/Calls. Thus if you never call, you will have an infinitely high AF.

→ if a player has a high AF and a high WTS, he plays very aggressively post-flop

From experience, loose fishes with a high VPIP pre-flop and a high WTS post-flop often have low AFs. Calling stations have an AF around 0.6. If you now face one of these loose players with an AF of 1.3 though you have to be careful as he is aggressive post-flop. At times he makes semi-bluffs and in some spots even pure bluffs. Thus you should check VPIP, AF and WTS in combination.

Examples: Which player type has such stats?

• (25/18/2.5/37) ....TAG
• (60/04/0.5/48) ....Calling station
• (60/04/1.2/45) ....Calling station, aggressive with draws and able to bluff
• (69/33/2.2/45) ....Maniac
• (35/25/2.4/44) ....LAG
• (35/10/1.1/41) ....semi-loose, weak-passive player

Conclusion

Stats are a most important tool. They enable the own game evaluation, and a better evaluation of the opponents. The own perception and observation are also important, but they are also always subjective.

Stats on the other hand are objective. You can persuade yourself subjectively that you would at times play a hand outside of the ORC, but when your VPIP is 33, then there is no way to put a gloss on it.

References were recommended in this article with which you can compare your own game. But despite the objectivity of the values, there's enough room for interpretation as the dynamic component of poker requires an adaptation to the different game textures. Thus you shouldn't try to change your own stats only to make them look better. You should rather look for fitting spots. The best way to find these is by discussing special hands, either in coachings, sample hand evaluations, videos or discussions.