Leveling wars in poker: Reverse tells, reverse-reverse tells, and more

Uncertainty and anxiety tend to make tricky behaviors rare

Some of the thoughts below I haven’t spent a whole lot of time thinking through, and I haven’t played much poker in the last few years, so all thoughts/feedback appreciated.

I was thinking recently about “leveling” in poker, which is what some people call it when you try to manipulate an opponent’s decision by doing something that you think will be interpreted a certain way.

For example, in the realm of tells/behavior, this might be a reverse tell: performing a behavior that you think your opponent will interpret as a common tell and lead them to make the wrong decision. Or it might be a reverse-reverse tell: switching it up another “level” beyond that. If you succeeded at this, you might say that you’d “leveled” an opponent: this is something occasionally said by poker players (and it may also be said in other games, I’m not sure).

This is related to what, in psychology, they call recursive mind-reading: the process by which someone tries to figure out “what he thinks that I think that he thinks…” and so on.

I was recently corresponding with a high-stakes poker player, talking about tells, and he asked me what I thought of this situation:

I played with a friend who knows about your work and we do a lot of leveling between us for fun. In the hand he made a significant river bet and, when I prompted him with a joke, he had a genuine smile. My interpretation was that he would not genuinely smile with a strong hand because he knows that I know it’s hard to fake with a bluff. I was wrong and he had a strong hand.

Later he told me that he chose to smile because he assumed that I wouldn’t think that he was foolish enough to do that with value, as we both know the common meaning (relaxed and strong).

So he leveled me here into thinking that he was trying to use a genuine smile with a bluff.

I’ve occasionally had people ask me similar questions: asking about how common reverse tells are; or how common reverse-reverse tells are. The practically useful question is: How might we try to deduce what “level” an opponent’s behavior is on?

Mapping out levels

In my book Verbal Poker Tells, I described levels this way, regarding statements that communicate something (directly or indirectly) about one’s hand strength:

• Level 0: The truth.

• Level 1: A deception. This would be the “weak means strong; strong means weak” paradigm, where most useful verbal tells reside.

• Level 3: The truth; reversing things for an opponent who expects Level 1 behavior (this would be a reverse verbal tell).

• Etc…

But that mapping was only meant to be for verbal statements about hand strength. The situation is much different if we’re trying to describe “leveling” in general, for all non-verbal and verbal tells. The reason is because there are many common tells that are indicators of someone’s genuine emotion/feeling. There are also many commons tells where

Recently I spent some time thinking about how you might map this out for tells in general:

• Level 1: The “common tells” level, where almost all common tells reside, whether a leak of genuine feeling or done more deceptively

• Level 2: Reverse tells: reversing the common meaning

• Level 3: Reverse-reverse

• Etc.

Level 1 is where almost all common and practically useful tells reside. Note that this categorization isn’t getting into the motives or reasoning behind why they’re done. This is a way to be able to include all types of tells; as stated, many tells leak genuine states/emotions (like smiling genuinely when betting a strong hand; or not being mentally focused early in a hand when you have weak/medium-strength hands), while many do not (most verbal tells are weak-means-strong and strong-means-weak). All that matters here is that they have a fairly predictable meaning amongst the general player pool.

Level 2 is what would be called “reverse tells.” An example that comes to mind: a skilled player bluffs and calls the clock on his opponent. He hopes that that player is proficient enough with tells to see his calling the clock as what it will usually mean: a sign of relaxation, and a strong hand.

Level 3 would be so-called “reverse-reverse” tells. Let’s say in that example hand just given, his opponent called and saw his bluff (or was shown after folding). Let’s say a year later that first player was facing the same opponent. He finds himself in a hand with this opponent and this time has a strong hand. He shoves all-in and calls the clock. Here he could be said to be performing a reverse-reverse tell; he is hoping his opponent thinks he’s doing a reverse tell (Level 2), like he did before, while he’s actually thinking on Level 3.

Reverse tells (and beyond) are rare

In my books and videos, I’ve talked about the general point that reverse tells (and above) are quite rare. Why are they so rare?

The primary reason is that, in a competitive environment where deception is assumed, it’s hard to predict how one’s actions will be interpreted. This is especially the case when it comes to tells, as usually there’s no easy way to tell what someone’s knowledge of tells is like. Even if you did know what an opponent’s knowledge of tells was like, if they in turn know that you’re a decent player and that you know something about tells, this sets the groundwork for all sorts of uncertainty about how things will go.

Note that all that is true for a reverse tell. It becomes even more confusing for a reverse-reverse tell (and above). You’ll be trapped in a world of confusion and second-guessing yourself. And this is why people don’t much attempt it. Just think about it for yourself for a bit: think about how you’d try to perform a reverse-reverse or reverse-reverse-reverse behavior versus an opponent. Or think about how you, as someone facing a behavior, would try to figure out if someone was reverse-reversing or reverse-reverse-reversing a tell?

Trying to think about what to do, and how it will be perceived, at higher levels just becomes madness; like trying to navigate a car while looking through a kaleidoscope.

Even just reverse tells (level 2) are much more rare than I used to assume they were. I used to study a lot of footage of high-stakes games, and I expected to run across a good number of “leveling” wars with some reverse tells in the mix; but no, they were extremely rare. Any interesting behaviors I found were mostly just more subtle versions of the most common tells. (Now, there’s a chance some of those common tells were actually level 3 reverse-reverse tells but I almost never got an impression there was much in the way of complex leveling going on.)

That said, such “leveling wars” become a bit more common when serious players play with each other often. But in general my understanding is that almost all serious regular players would strive for behaviorally GTO approaches: just trying to be stoic/unreadable in important situations.

The important thesis: Level 2+ behaviors from bettors will mostly be strong hands

But here’s the important part of this: When someone making a significant bet does attempt to get behaviorally tricky, they’ll usually have a strong hand. The reason for this has to with the uncertainty and anxiety just discussed: it’s hard to figure out how an opponent will react to a behavior. Therefore the people willing to try it at all will tend to be quite relaxed.

But there’s an additional reason here that adds to this pattern: bluffers don’t want to look stupid. Bluffing is already stressful enough. A bluffer doesn’t want to try some risky, uncertain strategy (e.g., a reverse tell like saying “I’m bluffing”), be called, and then have to deal with the feeling that his behavior induced a call. Better to keep stoic and quiet; better to stick to the usual script.

The fear of looking stupid plays a big role in such spots, in my opinion (and I think it plays a role in a lot of games and competitive situations). It constrains people’s behavior.

As a way to illustrate how powerful this is: Many serious players are aware that when a bettor weakens their hand range verbally (what I call a weak-hand statement), that will be highly correlated with a strong hand (more on that in this Upswing Poker article if you want to learn more). Despite the fact that many good players know this (I think in large part through me talking about it), you’ll almost never find a serious player making a big bluff while verbally weakening his hand range. Even though one would likely have very good results with it, it’s just extremely rare.

I can say for myself that, even if I thought my opponent was likely to read my weak-hand statement as strong, I’d still feel quite nervous doing that when bluffing. We have that fear of getting called and our opponent saying, “I called because of what you said.” Or we fear just thinking it was a factor in the call.

(I’ve focused on bettors here but what about a non-aggressor player: someone checking or calling? In general, any sort of unusual behaviors from them will tend to equal weakness, so that pattern will still hold.)

Here’s something I made to try to summarize the points just made:

Sometimes a tricky attempt at “leveling” will just be a standard tell

One upshot of the points I’ve made is that: Often, when a player betting a strong hand believes they are trying to be tricky and level someone, they may actually just be performing a normal, run-of-the-mill tell.

Let’s return to The Case of the Genuine Smile from the start of this post. The hero’s opponent bet and performed a genuine smile. Here’s what that genuine smile would represent at the different levels:

• Level 1: Value-bet

• Level 2: Bluff

• Level 3: Value-bet

An important question here would be: How likely is it that this player would try to display a genuine smile when bluffing? Unlike some other tells of strength he might have displayed, a genuine smile can be hard to pull off. It’s hard to fake a genuine smile (affecting eyes, dynamic, etc.). It’s hard to know what to do with it once you’ve put it on (let it fade away? try to keep it up?)

This player said that he “chose to smile because he assumed that [his opponent] wouldn’t think that he was foolish enough to do that with value,” which implies Level 3. But I think it’s worth asking: Was this player really going to Level 3? Or did his relaxation just lead him to performing a behavior that he would seldom perform with a bluff — regardless of his stated internal reasoning or explanation? Was it in fact a Level 1 tell, not a Level 3 tell? Or can maybe some Level 3 tells actually be not much different in practice than Level 1 tells?

Leaving aside the complex philosophizing, I think the practical takeaway is that: If you think an opponent is capable of trying to level you, and they’re making a big bet, and they have an unusual behavior, I’d tend to assign it the usual meaning — and most useful tells from players making big bets are tells of relaxation/strength. When in doubt, without some other reasons/history to refer to, it’s good to just apply Level 1 interpretations to tells.

Another way to put this is: Bluffers will seldom display Level 2 reverse tells. As my high-stakes acquaintance said about that hand: “Before thinking he might put on a genuine smile with a bluff, I probably should have first saw evidence that he was willing to do such things before.” Another way to say that is: Before thinking someone is capable of displaying a Level 2 deception with a bluff, you’ll preferably have want to have seen evidence that they’re capable of that.

A caveat: Not all tells are equal

A final caveat: some behaviors are more reliable tells than others, and some behaviors are easier for bluffers to fake than others. For example, “the speech” (talking a lot before or after making a bet) is known for being correlated with a strong hand, but it’s also not that reliable a tell. Bluffers sometimes ramble verbally, too; out of an attempt to reverse the well known tell, or out of nervousness, or whatever. On the other end of the spectrum, things like genuine smiles or weak-hand statements will be more rare from bluffers.

This is just to say that, as with all tell-related info, the ideas in this post are general ones. I think my thesis will largely hold for the more reliable tells from players making significant bets; the less reliable a tell is, the less reliable the info discussed here would be.

Another caveat: often people just do weird, random stuff

One more caveat: it’s true that sometimes people just do weird, random stuff. A lot of the more unusual behavior, especially verbal behavior, falls into that category. Sometimes you may wonder if someone who has those tendencies is trying to do some fancy, reverse-tell psychology on you. But sometimes it’s just people who want to be weird, in a “I’m a mad genius”-type way.

Still, having said that, there are some things that remain pretty unlikely even for people who try to act weird and random (e.g., making a weak-hand statement when bluffing). But for more weird-acting players, a lot will depend on what you’ve seen them do in the past.

When’s a good time to try to reverse a tell?

You might be wondering: “So when is it good to attempt a reverse tell?” In Reading Poker Tells I mapped out some of those reasons one would attempt it. An optimal situation in which to use false tells, to me, would look like this:

When you’re playing a player who you have reason to believe a) knows about the common tells, and b) would be likely to assume you’re a recreational (not skilled) player.

I’ve occasionally used false tells when a few stars align, but I tend to avoid it, for reasons mentioned (cognitive load, uncertainty, stress).

My high-stakes acquaintance and his friend know that their choice to engage in “leveling wars” isn’t optimal: they’re having fun with each other, and maybe learning some things about deception and trickery in the process. Just to say: there can be other reasons to try to get trick; really it’s all about how comfortable you are with such things and how confident you are in your predictions of what your opponent will do.

Have any thoughts/feedback on this? Love any stories you have about reverse tells you’ve tried, or that have been tried on you. If you don’t want to leave a comment, use my contact page. If you have an interesting story or thought, I might add it to this post or a later one.

Fear of looking stupid in other conflict situations

I work full-time these days on the problem of political polarization (my Substack). I think the fear of looking/feeling stupid plays a big role in how those things play out, and I think that aspect is barely discussed (just as we tend to underestimate the emotional aspects of our politics-related behaviors in general).

For political conflict (and other conflict), the way this would play out would be this: even if one believed that behaving in a more respectful, bridge-building way with an opponent would likely yield the best outcomes (for the entire group and for one’s own goals), one still must contend with fears of a) being wrong, or b) being seen to be wrong. One must contend with the fear that one is being taken advantage of and will feel like or be seen as a sucker.

And then, on top of that, there is the question of the time scale: even if you are quite certain a more bridge-building approach is optimal, how long until those benefits arrive? One might theoretically do what is the “right thing” in the long term for everyone, but still be seen in the short-term to be a sucker.

Optimal approaches to complex, conflict-related situations will sometimes feel wrong and counterintuitive to us. (Just as in poker, it might be a very effective strategy to make weak-hand statements when bluffing even as doing that feels very wrong and dangerous to us.)