Continuation betting (c-betting) is now so trivial that even the most inexperienced tournament players do it without thinking twice.
Today I hope to answer these questions:
- Why did near-automatic c-bets become the norm?
- What does frequent c-betting actually achieve?
- How should you go about c-betting?
We'll take a dive into specific c-betting scenarios and the math behind them later. But first, I want to talk about how c-betting has changed throughout the years so we can learn from past trends.
The evolution of c-betting
I've played tournaments professionally for almost a decade, and during that time I’ve seen c-betting strategy evolve. However, about six months ago, having begun studying game theory with Piosolver, I found myself clueless about the “why” behind certain c-betting decisions.
It was a humbling realization—despite playing probably five million hands in my lifetime, I realized that I didn't really understand why I was c-betting in certain situations. But nowadays I do. And so should you. So let's start from the beginning.
2008-2009: When you couldn't go wrong with a c-bet
When I began playing MTTs, around 2008–09, c-betting 100% was a powerful strategy. All the good players were doing it, and players not up-to-date with the magical new trend were getting owned—they’d just fold every flop they hadn't hit. So, c-betting 100% of the time made sense mathematically.
Suppose, for example, that we're c-betting using a generic half-pot sizing into a 100 chip pot. Since we are risking 50 chips to win 100 chips--2 to 1 pot odds--a continuation bet only has to work 33% of the time to return an immediate profit. Thus our opponent would have to call at least 66% of the time to prevent our c-bet from auto-profiting.
Since you only hit a pair or better on the flop around 33% the time in Texas Hold'em, the odds were heavily on the bettor's favor. Even if an opponent never folded a pair and continued with every reasonable draw, he'd still miss the flop and be forced to fold around 50% of the time--well short of the necessary 66%.
So, c-betting nearly every flop—in heads-up pots, at least—became the norm. It took years for opposing players to realize that maybe that c-betting 100% was a strategy they could exploit. (This seems hilarious to me now, at a time when trends in the game come and go so quickly.)
Back in 2011, I absolutely destroyed these frequent c-bettors with lots of raises. The thing that habitual c-bettors failed to think about was that, since we're all playing the same game, the c-bettor only hits a pair on the flop around 33% of the time as well.
(Note: Ready to take your poker game to the next level? Start crushing your tournament and cash game competition with expert strategies when you join the Upswing Lab. Learn more now!)
Furthermore, most players were c-betting not 100%, but more like 90%. And the only hands they weren’t c-betting were strong made hands--like sets--which they checked-back for deception, and some weak made hands for pot control. Thus, one would often face a situation like this as the caller:
30BB stacks, donkament.
Flop (6BB) T♦ 8♠ 3♥
Hero checks. BTN bets 3BB.
The button's c-betting range includes all of his air, and you've seen him check back nutted hands as well as middle pair. If we eliminate middle pair, then his c-betting range contains any pair only around a quarter of the time (see the range breakdown statistics on the right):
And since most players also folded any pair worse than top pair versus a raise, this c-bettor will have a made hand that can withstand pressure only 17% of the time. It's obvious that exploiting such a strategy with frequent raises is easy.
And what's best, check-raising with short tournaments stacks is both cheap and regularly forces your opponent to commit to the pot or give up. Consider again the previous example:
30BB stacks in a tournament.
Hero is dealt J♠ 7♠ in the big blind
Button raises (with an estimated range of ~50%) to 3BB. sb folds. Hero calls.
Flop (6BB) T♦ 8♠ 3♥
Hero checks. BTN bets 3BB. Hero raises to 8BB. btn folds.
We're risking 8BB to win 9BB, meaning that the check-raise has to work 47% of the time to return a profit (not accounting for the times we get called and go on to win the pot). Against someone using a 100% c-betting strategy—or, even better, a strategy that includes checking back some made hands but betting all of their air—it's now us exploiting the shit out of the guy who thinks he's exploiting us.
Nevertheless, the trend of c-betting near 100% went on for a long time. As late as 2012, major training sites were full of videos advocating c-betting close to 100%. And, worse, the videos usually ignored the math and logic behind c-betting.
Instructors would fire c-bets automatically, like robots, and only start with analysis after getting check-raised or when the hand proceeded to the turn. And the analysis was always the same: “I don't have anything, so I have to fold now”. These instructors were getting paid hundreds of dollars per video, and I was one of them!
I had made a lot of money exploiting automatic c-bettors, yet I lacked a sophisticated c-betting strategy myself. Of course, it barely mattered at the time—most fields weren't capable of exploiting over-c-betting, meaning that mindlessly c-betting was still very profitable. Why fix what ain't broken, right?
Then, around 2013, the tide started to turn.
2013: When players started looking at the flop before c-betting
Players were slowly, but steadily, catching on. Maybe c-betting A♠ K♠ on 9♥ 8♥ 7♥ wasn't such a great idea.
The over-c-betting population started board-selecting a bit; giving up on the worst boards, and paying more attention to their opponents’ tendencies. Opponents started playing back a little bit more, and attacking in spots where the c-bettors were obviously over-bluffing. None of the strategies in play were particularly sophisticated, but times were slowly changing. Still, the players who c-bet nearly every hand weren't done yet. It was time for the next revolution in continuation betting.
When it became clear that firing half-pot sized c-bets no matter the board was burning money, the brain trust came up with a new adjustment: using tiny c-bet sizings around 33% pot. Small c-bet sizings remain popular today, and it’s a pretty sensible strategy.
One huge upside is that a 33% pot c-bet only has to work 25% of the time to return an immediate profit. In order to prevent being exploited by a 33% c-bet, you'd have to play back at a 75% frequency while only making a pair on the flop 33% of the time.
The genius behind small sizing is that it’s not intended to fold out two overcards, ace high or gutshots. Instead, it aims to force folds from a small part of an opponent’s range that has completely missed the flop. Since it only had to work a small percentage of the time, how could it be bad?
Predictably, players eventually started catching up again.
The thing is, check-raise bluffing is very cheap when the c-bet size is small. With tournament stacks typically hovering around 30 big blinds, you can still force your opponent to either commit or fold without risking more than a few big blinds. Consider again our hand example:
30BB stacks in a tournament.
Hero is dealt two cards in the big blind
Button raises (with an estimated range of ~50%) to 3BB. sb folds. Hero calls.
Flop (6BB) T♦ 8♠ 3♥
Hero checks. BTN bets 2BB. Hero raises to 6BB.
This check-raise only has to work 43% of the time to return a profit (risking 6BB to win 8BB). Yet, the button will have top pair or better less than 20% of the time. Of course, he can also have all kinds of draws, which aren't folding. But he's going to be forced to fold a lot with a range that's at least half air.
Midway through the decade, both c-betting and check-raising tendencies were all over the place. Everyone was out to exploit everyone, and tournaments became a game of counter-exploiting the counter-exploiters. Players were taking strong lines in spots where they were representing very little, such as 3-bet shoving with total air over a check-raise.
Think about the above hand example one more time: what did our check-raise represent on that board? Random two pairs like 83 and T3 weren't in our pre-flop range, and strong hands like TT, 88 and T8 would be slow-played by most players.
Since no one was representing much despite bets and check-raises, the average pot size went up, and variance increased dramatically. For my part, along with a large chunk of the poker population, I was 3-bet shoving A3 on T-8-3 without much thought. Very often an opponent's check-raising range was full of shit, and if I happened to get called, I usually had five outs. No big deal, right?
Eventually, the tournament population got tired of playing bloated pots with weak hands, and the variance became intolerable. Something had to change. And so it did. Again. But this time with the help of game theory.
(Note: Looking to increase your confidence and earnings at the table? Master both fundamental theories and situational exploits when you join the Upswing Lab. Learn more now!)
2015: When players started to consider their own ranges
Game theory optimal play (GTO) became popular in tournaments around 2015, and it remains the most popular topic of study today. Unlike previous trends, playing a GTO-based game means playing non-exploitative poker.
Playing a perfect GTO strategy, there's nothing your opponent can do to exploit you, but neither are you exploiting anyone. It's far from a perfect strategy for tournaments, where most players are very exploitable.
Note that actually playing a perfect GTO strategy in No Limit Hold'em is impossible because it has yet to be discovered, even by computers.
A basic example would be a spot where the GTO play was to c-bet 40% of the time, but our opponent is a huge nit who massively over-folds to c-bets. C-betting with just 40% of hands would be leaving money on the table versus this player, no matter what GTO dictates.
Now, this could change soon if the quality of play continues to improve, but for now it’s safe to say that GTO play is suboptimal in most MTT scenarios. However, GTO solvers, like Piosolver, are great tools for figuring out why we're doing what we're doing.
Modern day c-betting
I downloaded Piosolver early this year. Having stood witness to many changes in c-betting strategy, it was eye-opening to finally gain a mathematical and GTO perspective on those changes. In particular, it was only this year that I finally realized the genius behind a 33% pot c-betting strategy.
Sure, the mathematical basis was still sound, but we can't expect to make an exploit last forever. Players will catch on, and sooner or later it will come back to bite us in the ass, as most exploitative plays eventually do. Solvers finally provide a theoretically sound answer to why c-betting small and often is the right play in many scenarios.
All of the following Piosolver simulations make the same assumptions. 30BB stacks, we open-raise with this 23% range:
...and we get called by the big blind with an estimated 42% range:
Let's start with a pretty obviously good c-bet spot, 7♥ 2♠ 2♦:
Even though we few top pairs in our range, we should fire a 1/3 pot c-bet most of the time. This is because our range is still stronger than our opponents:
Aside from the c-bet working out mathematically, we have other reasons to c-bet:
- Range advantage. Since our range is 58% against our opponent's 42%, we’re essentially making a tiny value bet with our range.
- Equity denial. Most of villain's hands still have six outs twice against most of our hands. So, even though we have a range advantage, most of our hands don't want to let villain realize their equity.
Note that these are both purely theoretical standpoints that don't account for other factors, such as ensuring we don't get bluffed out of the pot after checking back.
You probably already knew that you should c-bet 7-2-2 most of the time, but now you know why it makes sense to do so.
Let's look at another example: Q♥ J♥ 3♦. This is a flop where we only have top pair or better less than 27% of the time:
Yet, according to Piosolver, we should still fire a 1/3 pot c-bet nearly 100% of the time:
The same principle applies here, too. Even though much of our range has completely missed, and despite having an ace-high-heavy range on a board with two high cards, we're still way ahead of our opponent's range.
We have over pairs and top pair top kicker, whereas he doesn't; our middle pairs are stronger; our air has an over card higher than most of our opponent's air; and so on. With a range advantage, we still get to c-bet small nearly every time.
Finally, let's peek at a flop that's less ideal for c-betting: 8♥ 6♦ 5♠:
We only get to c-bet 38% of the time. This "heat map" shows why:
As you can see, most of our holdings are mediocre hands that don't want to build a big pot, and thus our range mostly wants to control the size of the pot. For every hand that might want to get it in, our opponent has a stronger counterpart. We might have AA, for example, but our opponent can have a bunch of two pair combinations.
And since we're a small dog overall, our range can't c-bet for value, either. Equity denial is less of a thing, since many of our hands that would benefit from equity denial have to either fold to a raise or get it in very bad against a check-raise.
If you use solvers, I recommend running these simulations on a regular basis. While there are many spots that are automatic c-bets, unpacking the reasons for c-betting in those spots is a great way to improve your poker thinking.
A three-step guide for improving your c-betting with solvers:
1. Know basic c-betting math. For example, a 50% pot c-bet has to work 33% of the time to break even, a 33% pot c-bet has to work 25% of the time to break even, etc.
2. Input both your and your opponent's range into a program like Piosolver. Find an answer to these questions: Who has a range advantage? How does the majority of my range want to play this texture? What percentage does game theory think I should c-bet?
From (1) and (2), sometimes you have a very clear answer. On boards that you smash, for example, a c-bet is justified mathematically and theoretically. In some cases, however, you won’t have a clear answer. When you get conflicting results—i.e., when the math says a c-bet has to work some percentage of the time that’s in conflict with what GTO dictates—it's time for step three:
3. Play poker. Consider your opponent's playing style, how he reacts to c-bets, how you can exploit him, etc.
In many c-betting scenarios, there's an exploitative adjustment to be made. But with an increased understanding of how things work, those adjustments usually need to be less dramatic than c-betting 0% or 100%. Deviating just slightly from game theory in the direction that you think is best should do the trick.
Uncertainty, to some extent, is what makes poker so enjoyable—it would be no fun if every situation was solved to perfection. But knowing why you're doing what you're doing makes it a lot more profitable.
Have any questions? Hit me up in the comments below, or on Twitter @chuckbasspoker.
Note: Ready to take your poker game to the next level? Crush your competition with strategies from world-class players when you join the Upswing Lab. Learn more now!
Read more from Miikka Anttonen on Upswing Poker: