Opponent Profiling - Making Use of your Statistics

Even if you had hand histories for every hand ever played by an online opponent, you probably wouldn't be able to make much use of the data as it stands. It would take hours of poring through the hands to even begin to get an idea of the type of player that he or she is. And since you probably don't play the same few players every time you sit down, you would have to do this for thousands of sets of histories.

Fortunately you don't have to do this. Software programs have been written specifically for the analysis of hand history data. Two of these support a wide range of sites and have many users: Poker Tracker (PT) and Poker Office (PO). Which of these you choose is up to you, though some sites are supported by one but not the other (for instance at the time of writing only Poker Office supports iPoker Network, while only Poker Tracker supports WPEX). Threads regarding their comparative merits are commonly found in the relevant forums on Two Plus Two and other online poker forums

These software programs use the hand histories you supply them with to calculate several summary statistics about each player (listed in the table below). In the remainder of this article, I will discuss the meaning and significance of each statistic in more detail, and begin to give you some idea how you can make use of the knowledge you will gain about your opponents from their statistics...

The Key Statistics

Statistic	Example	Meaning of Statistic
VPIP	24.6	Voluntarily Put Chips In Pot %: Percentage of hands in which a player puts money (other than the blinds) into the pot
PFR	7.5	Pre-flop Raise %: Percentage of hands in which a player raises the pot before the flop
PFA	2.3	Post-flop Aggression Factor: A measure of how aggressive a player is after the flop
WTSD	31.4	Went To Showdown %: Percentage of hands in which a player goes to showdown having seen the flop
W$SD	48.5	Won at Showdown %: Percentage of showdowns won by the player
BB Steal	56.8	Blind Steal %: Percentage of hands in which a player has an opportunity to steal the blinds, and attempts to do so (by raising preflop)
BB Defence	66.6	Blind Defence %: Percentage of big blind steal attempts which fail against this player (because they called or reraised the 'stealing' raiser)
Continuation Bet	83.5	Continuation Bet %: Percentage of the hands in which the player raised preflop then bet out on the flop

The remainder of the article will deal with these 8 statistics in more detail, and also touch upon a couple of sets of statistics that are in general more difficult to apply and/or less important than those mentioned above, but are useful in certain situations. These are Check Raise % by street and Fold % by street. First, though, I will begin with by some distance the most useful, most quoted statistic, as well as that which requires the fewest hands to give you a good idea of an opponent's basic style...

VPIP

VP$IP (voluntarily put chips into the pot) is the single most important statistic when it comes to making an informed 'read' on an opponent. Within 10-15 hands, VP$IP begins to give you an idea of their general playing style. Put simply, VP$IP is the percentage of hands with which the player is prepared to put money into the pot, either calling or raising before the flop. It usually includes the small blind position, but not the big blind unless they have had to call a raise in order to see the flop.

Optimal VP$IP differs between 6-max (short-handed) and full ring NL Hold 'Em. Typically, the most profitable full-ring players have VP$IP between 17 and 22. In full-ring, players with a VP$IP percentage greater than the upper bound of this range are regarded as fairly loose. A 6max (short-handed) player with a VP$IP% of 22, on the other hand, would be regarded as fairly tight. Optimal VP$IP is typically regarded as being between 26 and 30. The reason for this is that the impact of the blinds is much higher than in a typical full-ring game--they come around twice as often (every 5 hands rather than every 8 or 9).

Knowledge of an opponent's VP$IP statistic can give you a good indication of the range of hands that they play. Though the following is not foolproof, it is a guideline that you can use along with your data 'reads' on the HUD that you consult as you play. Particularly, bear in mind that 6max players' hand ranges should differ slightly from full-ring, as Aces and even Kings with medium kickers become more valuable hands at the expense of suited and connected cards.

VPIP Range	General Player Style	Typical Hand Range (Outside Blinds)
<10%	Rock/Set Miner	AA-77, AJ+, KQ (Some may play all pocket pairs and AK only)
10-15%	Tight	AA-77, 'Picture cards', A9+
15%-22%	Fairly Tight	Above cards plus lower pocket pairs, higher connectors (e.g. JT), suited aces
22%-30%	Fairly Loose	Above cards plus all suited connectors, some unsuited connectors and one-gappers and some high cards (e.g KT, A8o) esp. in 6-max
>30%	Loose	Above cards plus unpredictable others

As you can see from the table, with just this one statistic you have some really valuable information about your opponent's pre-flop playing style. You can know, for instance, that a player with 11% VPIP, outside the blinds, is much less likely to have the straight on a flop of 5c 6s 8d than a player with VPIP of 36%. However, other than players with extremely high VPIP percentages (above 45%), what it doesn't tell you is how good the player is - whether in poker terminology they are a 'shark' or a 'fish'.

One way of differentiating between good and bad loose players is by consulting an associated statistic to VPIP, VPIP(LP). This is displayed in different ways by different tracking programs, but its purpose is to assess whether the player concerned is positionally aware. If this is relatively high compared to the VP$IP statistic, this is an indicator that the player is taking advantage of late position to enter the pot with holdings that they might lay down with more players still to act. Although this statistic requires data from many more hands to converge (become an accurate marker) than does VP$IP, it can be very useful as an indicator of the skill level of a loose player. Particularly at low stakes NL where loose players abound at many sites, this is invaluable.

PFR

PFR (Pre-flop Raise %) is the percentage of all hands that the player raises before the flop. In combination with VP$IP, it is a key indicator of a player's preflop style. It is generally regarded as the second most important player summary statistic. It can be exceptionally useful in both table selection (to be discussed in more detail in a future article), and in putting a player on hands in-game.
For an example of this, consider a player on whom you have a lot of data and has a PFR % of 1.8%. This would suggest that they raise with only 2 starting hands, AA and KK, and that your JJ is very likely to be behind.

A typical PFR % for winning players in full-ring NL is between 5 and 8 per cent. For 6-max (short-handed) players, this is quite a lot higher - tending to range between 10 and 12 - largely due to the greatly increased importance of blind-stealing. Some typical PFRs and their corresponding raising ranges are indicated in the table below

PFR Range	Typical Pre-flop Raising Hand Range
<3%	AA, KK, QQ, AK
3-6%	TT+, AJ+
6-10%	77+, A9+, KT+, perhaps some suited connectors
10-15%	Pocket pairs, picture cards, suited aces and connectors
>15%	Above hands plus unpredictable others

The above is a useful guideline for player reading in-game. Using your player tracking software (PT or PO), you can see exactly which hands a given opponent has raised pre-flop.
It can reasonably be expected that they will broadly conform to the above guidelines. If they do not (for instance, a player who only raises with pocket pairs), note it down in their player notes.

As with VP$IP, a little information about the degree of positional awareness displayed by the player goes a long way to assessing their level of ability - particularly, to what extent they solely "play their cards" rather than taking other factors into account. This can be adjudged by assessing whether the player's PFR % differs between early (UTG, MP1) and late (CO, Button) positions (you will need at least 1,000 hands on the player for this to be reliable).

An alternative way of assessing the role position plays in an opponent's game is by looking at their Blind Steal % statistic. As with VP$IP(LP), this should be relatively high compared to the PFR percentage if a player who is aggressive pre-flop is making the most of opportunities to win a(n albeit small) pot with a mediocre hand.

More than any other statistic, the value of PFR can be significantly increased with the addition of some extra information that cannot be summarised by a simple statistic, but can be found by spending a little time getting to know your player tracking software, and studying your most frequent opponents' play while away from the table. The particular pieces of information I suggest adding to your player notes are as follows:

To what amount of chips (how many times the big blind) do they typically raise with a given hand (for PL and NL)?
With which hands will they re-raise when someone has already put in a raise?
With which hands will they re-reraise when their original preflop raise has been reraised?

This information can be very useful. It can't be displayed by your HUD, but should be copied to your player notes at the earliest opportunity. Particularly at low stakes, players exist who can be put on exact hands or very narrow ranges of hands using the aforementioned information. Of course, other more skilful players will play the player a little more, and thus not act as predictably as this - varied ranges can act as an indicator of such 'sharks'.

PFA

PFA (Post-flop Aggression Factor) is calculated by dividing the number of times that a player bets or raises after the flop by the number of times they call. The post-flop aggression factor found in Poker Tracker averages over the three post-flop betting rounds - it is possible also to find aggression factor by street (AF(Flop), AF(Turn) and AF(River)). Of these statistics, the first is probably the most useful (I will discuss its application in more detail in another article). However, because the averaged statistic is most frequently utilised and quoted, I shall concentrate on this here.

In some circumstances it can be misleading, and for this reason is a little controversial, but PFA remains a good indicator of a player's general style:

A "passive" player (sometimes referred to as a 'fish', particularly those of the looser variety) has an AF of less than or equal to 1. This means they call as often as they bet or raise.
Some of these passive players can easily be pushed out of the pot (are 'weak-tight'). Others are 'calling stations' who will foil attempts to bluff them. It is important to differentiate between the two and this can be achieved by considering WTSD, the percentage of the time a player sees the showdown having seen the flop.

A slightly aggressive player has an AF of between 1 and 2. This is average, and provides limited information on the player's post-flop style without consulation of other statistics. Often, players with AF scores in this range, combined with VP$IP in the recommended range for the form of NL Hold 'Em being played, are solid, winning players--particularly in looser low stakes full-ring games where value-betting is a winning strategy and bluffing is typically not.

Players with AF scores greater than 2 are aggressive. It is likely that some of their bets and raises will be bluffs and/or semi-bluffs (aggressive play with a drawing hand such as a four-flush or open-ended straight draw. An opponent with an AF of 4 or greater combined with a relatively high PFR for the game can reasonably be described as a "maniac". These are the types of players for whom a bet or raise is the norm, and a call is perhaps something to be wary of, as it may well be an attempt at deceit with a big made hand.

The latter types of players can be difficult to play against, but particularly if they have high VPIP and/or PFR percentages you can typically reasonably consider top pair, particularly with a good kicker, to be in the lead despite the bets and raises coming in.

The Supporting Statistics

The three statistics described above are generally regarded as the key opponent 'reads' in online play. In forum discussions, poker data is typically quoted as follows: 24.6/7.5/2.3. This means that the player under discussion has a VP$IP of 24.6%, a PFR of 7.5%, and an aggression factor of 2.3. They are a slightly loose aggressive player, a style that with skilled post-flop play can be reasonably expected to be profitable in full-ring NL Hold 'Em.

However, they do not tell the full story about the style of play an opponent - in fact, they leave several gaps of knowledge, particularly post-flop. The next two statistics add a great deal of value in this latter area...

WTSD

WTSD (Went to Showdown %) is the percentage of hands in which a player sees a flop and continues through to showdown. There is a strong argument that as a single statistic WTSD is as useful, if not more, as VPIP when it comes to identifying poor players. (Though of course using both is better). A WTSD percentage of 20-25% is considered to be typical for a good player. Much higher than 30% and a player is likely to be seeing the showdown with hands they should have folded. Bluffing a player with a WTSD score of more than 35% is not recommended.

WTSD % by itself is a very broad indicator of post-flop playing tendencies. To really make the most of this statistic, it must be considered in combination with the previous 3 (and the following one, W$SD). I will attempt below to detail the most useful information about an opponent that can be gleaned from combining WTSD with other statistics:

High VPIP combined with high WTSD - indicates a particularly poor player who can be exploited with value bets and is unlikely to keep their chips for long (though they may get lucky in the short term)

High PFA combined with high WTSD - suggests that the player may not be making sufficiently large bets to push their opponents out of the pot, perhaps allowing other players to draw out against them

Low PFA combined with high WTSD - is the mark of a classic passive 'calling station' whose tendency to call can be exploited with value bets but should not be bluffed. They are unlikely to be a winning player

Combination of W$SD with WTSD will be discussed in detail in the next section

As should be beginning to become clear from the above, taking into account WTSD in-play can greatly assist you in targeting your opponents' weaknesses in-game.

W$SD

W$SD (Won at Showdown %) is the percentage of a player's showdowns that they won chips at. To a significant degree, you can reasonably expect this statistic to be inversely correlated with WTSD %. A typical won at showdown % for a player with an average WTSD (~25%) is around 50%. A player with a W$SD of less than 50% over a large sample size (over 10,000 hands(!)) is likely to have one or more leaks in their play.

Of course, apart from yourself, you are unlikely to have many players with more than 10,000 hands in your database. As such, I suggest that you take this statistic in particular with a heavy pinch of salt, as of all the statistics mentioned here it is the most heavily influenced by short-term luck. With fewer than around 1,000 hands, the statistic is not really useful.

To make use of W$SD, you are likely to be concentrating on players with 'extremes' of the statistic over 1,000 or more hands. The majority of players will have W$SD scores between 45% and 55%. If, however, a player's W$SD percentage is outside this range, it may be worth looking into the meaning of this as regards their style of play. To do this, consider WTSD and W$SD together, broadly as follows:

High WTSD combined with High W$SD - This is unlikely (recall, 'inversely correlated'). It can occur occasionally in the statistics of tight (low VPIP), passive players ('mice' or 'rocks') whose hands have been holding up.
High WTSD combined with Low W$SD - This on the other hand is common, particularly at lower stakes. Along with high VPIP this is the best indicator of a poor player ('fish'); combined with low VPIP it suggests a tight-passive player who has failed to avoid their premium hands being outdrawn.
Low WTSD combined with High W$SD - This is the commonest of the four combinations. The player is likely to be winning, but it can indicate that the player folds too easily if they suspect that they may be behind, even when they have correct odds to make the call (especially on the river)).
Low WTSD combined with Low W$SD - This is quite rare, but sometimes found in the case of aggressive players who play strong hands strongly (resolving the hand before showdown) and weak hands weakly (often checking down). It can also sometimes help identify players who bluff too weakly and are called, again particularly on the river. The river-specific Aggression Factor statistic can be used to back up (or contradict) the latter theory.

There is another statistic that you can use to add information to your post-flop reads in combination with the previous two, W$WSD (won chips when saw flop). This statistic requires as many hands, if not more, to converge as does W$SD and primarily for this reason I mention it here only for completeness.

Blind Steal %

Blind Steal % is the percentage of opportunities to steal the blinds that a player takes. Unlike the previously described statistics, this percentage indicates how often the player performs a very specific action - raised preflop from the button or cut-off seat when no-one has yet entered the pot. This is particularly important in 6-max; the described situation arises less frequently in full-ring.

Apart from helping you to decide whether or not you should 'defend' your blind (more on this in the next section), as hinted upon earlier the Blind Steal % statistic is in combination with PFR a good indicator of whether or not a player is positionally aware or is simply playing their cards. High PFR % combined with low Steal Blind % suggests the latter. Basically, this statistic is an indicator of how much you should respect late-position raises from the opponent concerned.

A good player's blind steal % should be at least 20%, based on the supposition that if you have a hand that is in the top 20% of starting hands it is unlikely that either of the blinds will have a better hand.

Blind Defence %

Blind Defence % is the percentage of the time that the player defends their big blind against a steal (calls or reraises the late-position raiser). You needn't take much notice of this statistic except for the two players to your left. The lower the percentage, the more frequently you should attempt to steal their blinds.
However, note that particularly at low stakes and in full-ring NL, this percentage may well justifiably be low, since blind stealing is quite rare in these games and raises are more likely to be from genuine raising hands).

Continuation Bet %

Continuation Bet % (sometimes referred to as PFR(Flop Bet)) is the percentage of the hands in which a player raises pre-flop then bets out before the flop (disregarding those occasions when a player acting before them on the flop bet first). Again it is a very specific statistic the knowledge of which can be exploited to take advantage of a player's tendency in this area.

Players who have a high continuation bet % along with anything but a very low (below around 6%) PFR are betting the flop with Ace-high or worse having raised pre-flop, and should be called or re-raised much more often than opponents with comparable PFR % but much lower continuation bet %. The latter group of players tend only to bet the flop when they have a hand - it is probably worth consulting their other statistics (particularly PFA and WTSD), to see if a player who rarely makes continuation bets is generally passive or aggressive. If the latter is the case, it suggests that they are broadly straightforward, honest players who tend to bet when they have a hand and fold when they don't. This is because continuation betting is one of the most common types of bluff (along with semi-bluffing), and a player who does not bluff in this situation probably does not bluff much elsewhere either.

Street-by-Street Statistics

Two other sets of statistics that have not been mentioned thus far are also available for use through player tracking software and HUDs. These are the flop, turn and river Check Raise % and the flop turn and river Fold %.

Check-raise % is a statistic that takes a long time to converge--at least 500 hands. It is an indicator of how 'tricky' a player is. A player who never check-raises has something of a leak in their game, as this means that a check from this player when first to act almost certainly indicates weakness. On the other hand, a check-raise % of 2 or greater on any street probably indicates a player who is overly tricky, probably afflicted by 'fancy play syndrome' (making supposedly 'clever' moves when straightforward value play would be more profitable). Look at this player's other statistics, and avoid betting into a check from them with a hand you would not call a raise with.

Fold % should ideally be used in combination with by-street aggression factor to identify a player's general tendencies on each street. It gives the percentage of the hands a player sees the flop with, that they fold on each street. It can be used in combination with VP$IP and WTSD to help you judge the likely success of a bluff against this player.