Stats (1) - Basics
by torytrae
| 1 |
Introduction
In this article- How stats can help you with your game
- Which problems you need to be aware of
- Different levels of stats and their structure
In every hand evaluation, in every consideration and on every street, stats play a very important role. That's why you should know what stats say and what they don't say about a player in a given hand.
What are stats?
Singular: The stat (abbreviation of "the statistic")The term stats refers to all statistical figures that you collect about your opponents in a poker database with the help of poker software. Such software may include the PokerStrategy.com's Elephant, Hold'em Manager or Pokertracker.
As statistical mean values, the scope of information they provide about a specific situation is somewhat limited, but they still often determine your reaction in a specific situation.
This article kicks off an article series about stats, which will deal with exactly these aspects and their correlations. You will learn which stats build on each other and how the interpretation of stats helps you with your decisions in the game.
In this series, you will learn about the basics of the pyramid principle and how the different stats relate to each other. Later we will focus specifically on the interpretation and correlations of individual stats as well as the influence of sample size. On top of that you will see what role stats play in your meta game.
You use stats in order to limit your opponent's range in a given situation and to figure out how the equity of your hand or of your range would hold up against it. Stats also help you get an idea of what player types you are facing and how they tend to play.
Stats are statistical mean values and you need to keep this in mind. Stats therefore tell you how an opponent, on average, reacts to a certain situation, e.g. how often he raises or calls preflop.
Stats predict tendencies. However, there are other factors that influence every decision your opponent makes. Looking at stats out of their context can lead to wrong decisions if you don't see or don't consider other pieces of information.
Stats may therefore not serve as the only basis for your decisions in the game. You have to embed their interpretation in the overall context of the situation. You should by no means blindly rely on their numerical value.
When you look at a stat, you combine two things: You take the stat's numerical value and based on this derive a range that your opponent might have.
In general, stats initially don't tell you anything about the strength of Villain's range, even though this may sound contradictory.
If, for example, your opponent has a 3-bet stat of 5%, this only tells you that this player has 3-bet 5% of the hands he has played. Nothing else.
Such a stat on its own does not initially allow any other conclusions about your opponent's range, but only represents the frequency of a specific reaction/situation. This is a very important aspect, which you need to internalise.
Stats specifically indicate the frequency of an action or a reaction. Stats do not directly indicate the strength of a range.
Every situation in poker - and this applies both to Villain and to you - depends on a number of factors. This includes playing style, emotional state, reads, history, other players in the hand, table dynamics, position, image, limit and much more. Since all these factors differ in every hand, they inevitably influence Villain's reaction in one way or another.
As you can already see, the decision in favour of a certain action or a reaction in a given situation depends on much more than just the hole cards you were dealt. This is true for your game and it is true for your opponents' game.
Stats as such are characterised by the fact that they always represent statistical figures out of their context, while they don't take into account other factors relevant to the situation, for example your meta game. This is another aspect you have to internalise.
One example for this is the "continuation bet flop" stat. This stat tells you how likely it is that Villain will make a continuation bet as the preflop aggressor. Of course Villain won't contibet on every board from every position and against any number of opponents. He will therefore base his decision on his "perceived strength", the number of opponents in the hand, the stats or the reads on his opponents as well as other considerations.
Looking at this stat out of context will of course not tell you if Villain often finds himself in a favourable situation on the flop. The stat does not tell you a thing about the conditions in which Villain made his contibet decisions, which in turn are the basis for the statistical value that you see.
When you deal with such a stat, you can therefore not assume that Villain will always contibet every board in every situation against any number of opponents at the percentage indicated by the stat. If you want to use a stat appropriately, you always have to take all the other factors into account as well.
Ideally, you should look at the stat in the overall context of the situation and adjust its validity accordingly. This is called 'interpreting' a stat. The concept of interpreting a stat is another important principle, which will be dealt with in further detail in the second part.
Villain has a contibet flop stat of 70%. If Villain is in position and hits a flop such as AQ7 rainbow against a preflop caller, he is likely to contibet in more than 70% of cases. If Villain is out of position, however, and he hits a flop such as 865 two tone against several opponents, he will contibet far less than 70% of hands.
The stat, adapted to the context of the situation, helps you make decisions in the further course of the hand. It is important to understand that you need to specifically adjust every stat to the given situation - you have to interpret it.
Recall the above example in which your opponent 3-bets in 5% of all cases. It is likely that you automatically assume that Villain 3-bets with the top 5% of his range. However, this assumption in itself is not correct. This is because poker is a game of incomplete information since you often don't see Villain's hole cards.
As you already know, the 3-bet stat only represents the frequency of this action, but it doesn't tell you how strong Villain's 3-betting range actually is.
This means that you only know that Villain 3-bets 5% of his hands preflop, but it doesn't necessarily have to be the best 5%. Maybe Villain tends to slow play KK+? Or he never 3-bets AK because he "won't hit anyway". Does he know the concept of light 3-bets? Does he try to isolate a specific player type with crap hands or does he attempt to win the hand preflop against certain opponents by creating fold equity?
Villain can also base his decision for a 3-bet on other factors, which have nothing to do with his hole cards. That's what you do as well - and you can assume that other players play by the same principles.
You see that you sometimes draw conclusions from a stat that are simply not part of the stat as such.
What is also important when dealing with stats is to distinguish between the size of a range, i.e. the number of hand combinations, and the strength of a range, i.e. the number of hand combinations that are strong depending on the action and the board. A small range does not necessarily mean it is a strong range and a big range is not automatically a weak range.
This is a paradox mistake. Poker players usually base their own stats on a sample size of many thousands of hands, but think they can make "accurate" estimations about Villain's range when they have a sample size of only few hands. This is a huge leak.
Stats and their levels
To make it easier for you to better understand stats and their correlation to each other, we will use a pyramid scheme to introduce them. We are making the following assumptions:
- The higher the level of a stat, the less often the situation it describes occurs. Stats that occur only rarely are at the top of the pyramid.
- The higher the level of a stat, the bigger the sample size has to be in order for you to be able to interpret it appropriately.
- Stats of a higher level build on stats of the lower level(s) either directly or indirectly.
This level scheme is based on the fact that some stats can only occur when lower stats have already occured.
One example: A flop continuation bet can only occur if Villain raised preflop and he is the preflop aggressor. That's why the contibet is directly based on the preflop raise.
Within the pyramid system, stats are referred to by the Roman number expressing their respective level. Example: I OR or II 3b.
Not all stats can be categorised into one specific level on the pyramid, but rather describe a general playing style or cover several levels.
The AF represents the ratio of aggressive moves (bet/raise) to passive moves (call). The respective formula is usually quite simple: (Bets + Raises / Calls). An AF of 1.0 means that a player approximately bets and raises as often as he calls. If the AF is smaller than 1, his passive moves dominate and if the AF is bigger than 1, his aggressive moves dominate.
The went-to-showdown stat tells you what percentage of hands Villain takes to the showdown when he is on the flop. This stat is aggression neutral and therefore doesn't distinguish whether Villain bets, calls down or checks behind. This means that when you interpret this stat, you also have to consider which action he chose and how big his AF is.
This is possibly the most important, yet most neglected value there is. The sample size represents the number of hole cards that Villain has received. It doesn't matter whether Villain plays the hand or folds it preflop.
If you have a sample size of 100 hands for a 15/10 Villain, it means that this Villain has been dealt 100 starting hands and contributed money to the pot with 15 of them. He raised with 10 hands and with 5 hands he called a raise preflop. This already shows how little 100 hands actually say about Villain's calling range for example.
The sample size therefore plays a decisive role when dealing with stats, because it helps you estimate how representative a certain stat is. The higher the level of a stat, the more important the sample size becomes. As with everything in poker, the distribution of the hole card is subject to variance and becomes more accurate as the number of played hands in your data base goes up.
Stats of the first level represent the first possible reactions of an opponent and therefore don't build on any other stats. In general, you should have a sample size of at least 100 hands to make estimations about trends.
The VPIP is the number one basic stat. It tells you with what percentage of starting hands Villain is willing to put money in the pot. This includes limping, raising or calling. However, auto-posting the blinds is not taken into consideration here. This stat influences a very high number of other stats and serves as a basis for making estimations about Villain's postflop range.
The VPIP also entails the PFR stat, since a raise of course also represents voluntary contributions to the pot.
The PFR represents what percentage of starting hands Villain is willing to raise preflop. This stat includes open-raising, i.e. every first raise that is not a limp, as well as all other preflop raises (3-bet, 4-bet). It is a good indicator of preflop aggression.
The calling range is not a stat as such, but it is very easy to figure out. This range represents the percentage of hands with which Villain calls a bet preflop. It can be quickly calculated by subtracting the PFR from the VPIP.
The attempt-to-steal stat is the first specific preflop stat. This stat tells you what percentage of starting hands Villain open-raises from late position (CO, BU, SB). With almost every good player, this stat is higher than the overall PFR. This is because the later the player is to act and the fewer players are left to act after him, the bigger the open-raising frequency.
The fold-to-steal stat represents the frequency with which a player does not defend his blinds against a steal, but instead chooses to fold. This only includes situations in which Villain is in the blinds and up against an open-raise from late position. His reaction from the blinds - for example to a UTG open-raise - is not relevant for the fold-to-steal stat.
Stats of the second level build on stats of the first level. That's why you should always take into account the respective first level when interpreting these stats.
As stats of the second level occur less often in the game, a considerably bigger sample size is required. You should have more than 500 hands to make accurate predictions about Villain's tendencies.
The 3-bet is the first stat of the second level. It therefore builds on a previous action of the first level: the open-raise.
A 3-bet is therefore always a reaction to another action and can only occur when someone raised before Villain. This move consequently occurs less often and you need a bigger sample size for a meaningful interpretation.
- Correlation: I OR anyone – II 3B Villain
The fold to 3-bet / Call 3-bet stat heavily depends on the dynamics of the limit. The more aggression there is preflop, the more often Villain will be confronted with a 3-bet and the more often he has to make a decision. At passive limits, you usually need a bigger sample size than at more aggressive limits.
- Correlation: I OR Villain – II 3B anyone – II Fold/Call Villain
The Cbet stat represents the percentage with which Villain makes a continuation bet on the flop as the preflop aggressor. This stat is expressed in relation to the flop range and therefore builds on the I PFR first level stat.
- Correlation: I PFR Villain – I Call anyone – II Cbet Villain
The fold-to-continuationbet stat describes Villain's reaction to a continuation bet, i.e. situations in which Villain calls preflop and is confronted with a bet from the PFA on the flop.
- Correlation: I OR anyone – I Call Villain – II Cbet PFA – II Fold Villain
A donk bet is an opponent's bet on the flop against the preflop aggressor (PFA). The corresponding stat tells you with what frequency Villain donks on the flop. Also applies to re-raised pots.
- Correlation: I OR anyone – I Call Villain – II Donk bet Villain on the flop
The third level includes stats that occur even less frequently, since they build on previous stat levels. They only affect a small number of played hands in poker. You should have a sample size that comprises more than 1,000 of Villain's hands in order to make fairly accurate assumptions about his playing tendencies. With almost every stat of the third level, there is no limit to this sample size. You need to interpret these stats very precisely, which means you also need to consider stats of the lower levels as well as the current game situation.
This stat represents the frequency with which Villain raises the PFA continuation bet on the flop. This includes check/raises as well as raises in position. The stat can also be applied to re-raised pots. The raise-contibet stat is already reliable with a sample size of a few hundred hands if you are playing a very aggressive opponent.
- Correlation: I OR PFA – I Call Villain – II Cbet PFA – III Raise cbet Villain
A 4-bet is a reaction in the form of a raise to a preceding II 3-bet. A 4-bet is usually a strong move. In general you can assume: The smaller the limit and the more passive an opponent, the stronger you should interpret the 4-bet.
As a 4-bet can only be played in reaction to a I OR and a II 3-bet, this situation occurs only rarely. The more passively a limit is played preflop, the less often Villain gets the opportunity to make a 4-bet. The stat therefore depends on the dynamics of the given limit.
- Correlation: I OR anyone – II 3-bet anyone – III 4b Villain
A float generally refers to a move that seizes the initiative from the PFA. It only indicates the frequency of such a move. Conclusions about the hand strength and range can only be drawn if you also take other stats into account. A float is often played on the turn.
- Correlation: (Turn) I OR anyone – I Call Villain – II Cbet Flop PFA
The second barrel is the continuation of a flop contibet on the turn.
- Correlation: I OR Villain – I Call anyone – II Cbet Flop Villain – II Call anyone – III Turn Cbet Villain
Stats of the fourth level generally occur very rarely, because they depend on a rather lengthy chain of actions on the lower stat levels. You should generally have far more than 1,000 hands in your sample size to make reliable estimations about playing tendencies.
A third barrel is the continuation of a second barrel on the river. A third barrel can only be fired when the PFA bet both the flop and the turn.
- Correlation: I OR Villain – I Call anyone – II Cbet Flop Villain – II Call anyone – III Cbet Turn Villain – III Call anyone – IV Cbet River Villain
These stats refer to further possible reactions of Villain to a preceding III 4-bet. Depending on the stack size, Villain can call or ship (or fold).
Depending on the stack size, Villain will directly ship, 5-bet or fold in reaction to a 4-bet.
- Correlation: I OR anyone – II 3b Villain – III 4b anyone – IV 5b/ Ship / Fold
Summary
In this part of the series you have learned what stats actually are and what they can tell you about your opponent's game. In addition to that, you got an overview of the basic stats and learned about the pyramid structure of stats of different levels.
In the next part you will deepen your knowledge about stats, learn how to interpret them and how to approach the concept of sample size.
| 1 |
Next article:
Gold