There are times to check-raise and then are better times to donk. You should make your decision depending on how you want to manipulate your opponents range.

Here are situations where donking makes sense compared to check-raising:
- you don't want to inflate the pot
- you don't expect your opponent to c-bet often
- you do expect to get raised often and your hand does well as a bet / 3-bet
- Our hand is strong on the flop, but a bunch of turn and river cards will make it much weaker

( these are from matthew jandas "sub zero" 2p2, video on cardrunners).

I replied to this argument in post 36 of this thread. More or less, the flop changes who has the capped range.

Are you saying that there are hands that a caller can have but a raiser can't?

No. And although there's some possible pitfalls to my logic, I'm saying that even if the cold caller can't have AA in his preflop range. Even on a board of A72, while the cold callers range is capped at 77, he can still have a set or better more often than the preflop raiser (if his preflop range is much tighter). So even though the cold-callers range is capped, he has a higher frequency of the "nutted" part of the ranges between the two players (even if he has less combos). Therefore, the cold-caller has the option of starting the betting lead.

Check raising a wide range doesn't present the same problem as donking cause our range is entirely in tact. This will do wonders for our turn and river ranges as well.

Even if someone was capable of donking and c/r'ing with perfect balance and nut hands in both ranges, all the way to the river (which is extremely more difficult than you seem to think) you still miss pfr's c-bet so often. 2/3 pot 75% of the time is better than no c-bet all the times you donk.

He either gets to see info about your hand or he doesn't. When you have a donking range, he gets info, when you don't he doesn't. This will probably result in him checking back more often; but this is fine for us since we'll have a weaker range and we're oop, so I prefer a 2 street game.

Yes, there are times I like a donk but they are pretty rare. And they're almost always against a player(s) I can exploit or in multi way pots when c betting drops way off. If you honestly think 60/40 or 50/50 is the correct c/r to donk ratio then good luck to you.

I haven't seen this argument before and I have to admit it's very interesting. I was trying to test it, but ran into problems: let's fix the flop at A72. Assume the cutoff opens and the BB calls, and both of them have standard ranges. Who has the strongest range, and/or the most playable range? I'd be tempted to say that the big blind does, because he has a much tighter range than the CO opener. However, there's one major problem: I 3-bet AK, AQ, and sometimes even AJ, against a CO opener (depending on hoe loose he is). In fact, I just went over the hands I call in the BB against a BTN open, and there are very few: For all practical purpose, I play a 3bet/fold game from the BB against late-position raises. In fact, I player 3bet/fold even against early-position raises from very loose players, so it's more accurate to say that I play 3bet/fold against wide ranges. Against tight well-defined ranges I'd have a calling range, but in this case, I think on most flops the opener's range in this case would not be significantly weaker than my range.

I guess that the bottom line of this, for me, is that for me donking is not a big issue since I just don't play heads-up pots OOP frequently enough to bother with it. And, as we said, in multiway pots everyone agrees that donking has its places. If you, however, cold-call OOP often enough, then your range might be strong enough to actually be stronger than PFR's range on the flop. In that case, I guess it's definitely worth it to develop a donking range. In fact, it might be justified to donk 50% or even 70% of the time on certain boards, just like we cbet some boards very often, simply because your range hits them very well. I never thought of this, and now I see that I didn't think of this simply because it's not relevant for my game.

You do need to consider that there's a reason that many people play a game which is mostly 3bet/fold from the blinds: playing OOP sucks. You can say that your range is tighter, so card advantage will compensate for your positional disadvantage. But if your card advantage is so great, why not 3bet? I guess we can think of a hypothetical strategy where the hand is profitable when you don't 3bet (because of card advantage), but once you 3bet, villain folds exactly enough to make the hand unprofitable. In this case, it might be +EV to call with these hands rather than 3betting (you only need to consider the implications of this on your 3-betting strategy, since now your 3bet range is stronger and you have to compensate by adopting a polarized 3bet bluff range, but nevermind for now). So, yes, it's possible that a good strategy can be formulated that cold-calls with good hands because they are only profitable vs villain's weak range, not vs a stronger range. Theoretically, I don't know how to reply to that. This might be especially true in deep-stacked play, where 3-betting OOP does not have a big enough effect on the SPR to make playing OOP easier. However, pragmatically, I doubt that this is worth the hassle.

To summarize: I was wrong. A player that cold-calls often in the blinds can have a range that is stronger and more playable than that of the PFR. For such a player, adopting a wide and aggressive donk-betting strategy can be quite profitable (just like cbetting is a part of everyone's game, so should donk-betting be a part of this person's game). Donk-betting will probably never be as profitable as cbetting, because of positional advantage, but I definitely want to think of this some more. I might even go as far as playing a few sessions without 3-betting from the blinds, just calling, and then donking around 70% (based on the board) and seeing how it works.

A72 is not a great example of board where the cold-caller can have a strongest range (only 2 possible set, almost 0 two pair, raiser have a lot of TP), but there are situation/board where the caller should have a stronger range, so maybe think of doing a donk-bet.
For start take any board where as original raiser you do only few cbet (eg. J98 with FD), these spot can be great spot where a donk-bet can be optimal as the cold caller.

Two very good point made by RainbowBright are these:
-the important thing is the frequency of good combo, not the number of combo (eg. in a range of 5% 3 combo weight a lot more than in a range of 20%)
-by never donk-betting we allow the raiser preflop to play over 4 card instead of 3, this is a huge advantage (and he is choosing if build the pot or not)

And I don't see a great point on saying "player cbet too much, so c/r is better, so I never donk-bet". Can be true that against most villain they do too much cbet, so it's best to exploit this leak, but sometimes againt a good opponent donk-bet should be the right move. (and a lot of player badly react to donk-bet, so they can be exploited)

A suggestion can be choose a specific situation/board (eg. CO vs BB, CO range: xxx, BB range: xxx), and trying to study how the optimal play should look like on that board.

The MoP example only deals with the case where an early position player makes a small raise and the big blind calls. In that scenario I think that the correct cbet frequency could very well be very close to 100%. It's a completely different situation if the cutoff or button makes a 3x open and the small blind calls.

I propose that we start using the terminology "hand distributions" and "ranges" separately. Hand distribution to mean all the hands a player holds at the begining of a street prior to any action from either player. And "ranges" refer to hands which make up a particular action a player takes such as bet, check, call, etc.

I only make the suggestion because I don't know what you mean above. Can you explain it a different way and possibly specify the advantage. For example, I'm not sure what you mean by "our range is entirely in tact" and I'm not sure why that's so beneficial.

This is true, but you do seem to ignore all the times that he gets to check-back and see a free card.

I disagree here. I would mainly prefer donking vs check-calling only when I have a hand which I think I can get 3 streets of value from. So in these case, I do not want to play a 2 street game because I'm losing value by letting him check-back the flop.

I also disagree with the idea that always check-calling doesn't reveal information to him preflop. It tells him that he will be able to see 4 cards for free which he can then choose to change his preflop strategy. For example, hands like small suited connectors realize their equity alot over the turn and river, so it's very important for them to be able to see those cards, so they can realize if they have hit the flush or picked up their back door draw. And if the Villain can be confident that he can always see a free turn card, then he might be able to exploit us preflop by opening up more of the middle suited connectors then what should be profitable from a GTO sense. So I'm not saying that this is actually the case, I can't really prove it, but I just wanted to try and illustrate a way in which checking 100% of the time provides information for the villain which he can use to exploit our tendency.


Is there a reason you didn't address my list of situations and instead focused on numbers that I said were more made up, but at least are a starting part compared to statements like "pretty rare". I didn't think that the situations are that rare.

I also don't think that donking would be limited to a strategy against weak players. In fact, I can only imagine that it will have some if not small role in a GTO strategy so that actually not using it vs a good player will probably be more problematic than not using it against weak players where many different types of blunt tools are effective.

Actually from MoP, I remember some statements about how having a "strategic option" like donking can never actually lower your value for the game. In other words, as they solved the toy games, they found that as they gave the players more tools, such as check-raising and donking that the value went up for the player every time. This was always because the player could obviously ignore using the new strategic option which is more or less what you're saying here, but I thought I would just throw it in as something to think about.

Our range in tact means we can still hold any 2 in our distribution when he c bets or checks behind and it's the turn/river. Its guess it's possible to achieve this with a perfect donk and c /r balance, but......

I'm not ignoring that he can check back. If he checks 25 and we miss 66 of the time, then 18 of the time we're pretty happy withheld free turn as well. So 7 of the time we would have rather seen a bet go in and it doesn't. All the times we get his c bet with air,when we wanted to donk will easily make up for this.

It's tough for me to quantify "pretty rare" cause in FL my donk is zero. My estimate might be 8 or 9:1 for c /r vs donks in hu pots. And vs a very strong player it might be higher. (cause I'm not confident I can balance the donks and my other ranges without leaving some loose ends)

I'm not intentionally missing any points, I'm responding from a tablet and just comment on what I remember. (which iswhy there's no percent symbols above)

For someone to exploit us for never donking they would do whatyou said. (raise for speculative hands). But for this to happen, they'd need a huge sample, and have to notice that our flop donk is 1 instead of 5. Then they'd have to hope that no one between him and us are going to play. Plus, we could counter exploit by calling wider ourselves knowing we can benefit from a free turn as well.

The occasions where hedoesnt have a hand strong enough to value c bet, but we could have gotten 3 streets of value by donking are extremely rare. I'm sure you can estimate a number. Plus, when we c /r we can get 4 streets of value!

I have been playing ALOT since this thread and almost entirely online, mostly HU NLH and its so strange to see how my perspective has changed.

I am 33 and have been a serious student of poker since junior high, but my how the game has changed. I thank 2 + 2 for amping my game up more in a couple months than YEARS of reading and playing was able to do. Thanks all!

What's really interesting for me is this;
When I started this thread I didn't really agree with the "never donk bet" rule. I thought it was usually good to x to the raiser but i didn't understand why, it was just what i'd read in my LIBRARY of books so its what i usually did.
Now I'm teaching my wife to play good, winning poker online about every night and I am SO hard on her when she donk bets. It totally drives me nuts. I hate it! I wish she would listen and just NEVER donk bet again for the rest of her life. But I have come full circle in my own play. Even though I understand much better now why most people don't like the donk bet, I have recently started to find more and more spots to do it. Its like a big circle.

I think the main problem with donk betting, and the reason I don't want my wife doing it is this; (especially for my less experienced wife)

When you lead into the pfr and the pfr reraises on the flop its so hard to determine what he's repping. If you have little or no history with the player you just have no idea what his raise means. He may be raising because he has equity, he may be making a random bluff, his range could be polarized, or not, HE MIGHT JUST BE RAISING YOU BECAUSE YOU DONK BET! That's the thing that gets my wife into so much trouble. She's always donk betting with middle pair or say a pair of sixes on K72/K52 and she's always getting raised (at least called)and then she's in a jam.
I think that is a big problem with donking, aside from all the obvious stuff; just that you can't be certain if your opponent is repping something or simply raising because you donk bet. I do it myself, and it works well at NLH50.

Anyway, thanks for the help with this concept guys. It meant more this time i read it than it did the first time. Especially Rainbow Bright. I have been in so many conversations with you where I just COULD NOT TELL if you were SUPER smart or SUPER not smart. I think I just made up my mind. You are super smart but you like to "take the road less traveled" and i think you play devils advocate. Or you just have a really unusual way to approach the game.

I would LOVE to see you play. I just have no idea how your well thought out but controversial ideas would work in a real game.
L8r all

Aw man...I'm sad I missed this thread.
I had this exact same question several months ago and wrote a research article + tested it in my game in different spots.

I don't want to cross-post my article for SEO reasons, but I really think this was a big eye opener for me.

The general thesis was one of my favorite repeating ones in poker -- any line can be optimal given the right circumstances, we just need to understand why we are doing what we are doing and apply that information to make the best possible decision.

Reasons Donk-Betting can be worse than CR:
- If your range forks and your opponent is capable of exploiting you.
- Villain with a high C-Bet%
- Board texture impacts villain's C-bet% and structure, so just be mindful that Wet boards are likely worse spots to donkbet than Dry boards, in general.

Reasons CRing can be better than Donk-Betting
- If villain is braindead and can't exploit you
- Villain has a low C-Bet%
- Dry boards are generally better than wet boards

Basically, the issue with donkbetting is that most people do it with a highly imbalanced range and without much purpose. If there is no real purpose, then they haven't weighed the option against the options of CRing or CCing...they just chose the first +EV option they could think of (assuming it's +EV). If that's the case, then they're obviously not playing optimally, and likely aren't gaining anything but risking exploitation.

The mistakes people make are generally to only donkbet as a bluff or only donkbet for value in certain spots. This is okay against braindead people. Against an opponent with a brain, you'll find that they either consistently raise your donkbet or consistently fold to your donkbet, dependent on how your range veers from balance.

However, deviation from equilibrium is how we exploit people. If someone has a very low C-bet% and I have a value hand, then I don't really want to miss a bet. I can choose to polarize my range and donk with my value here and bluff on occasion, but I can also only donkbet for value against stations or polarize my range and weight it to air if they fold an exploitable amount of the time. Granted, if they had a high C-Bet% and it's imbalanced, I'd rather just CR and abuse the imbalance in their C-Bet range in a bigger pot...I stand to make a lot more money by CRing than I do by donk betting.

That's a pretty watered-down version, but I hope it helps!

Generally speaking, I don't donkbet much. In general, people who raise a lot tend to C-bet a lot. I'm only really exploiting people with donkbets when they raise a lot but have a low C-Bet%. It's just important to remember that there are definitely times when it's optimal to donkbet instead.

Only thing I disagree with is that donking is better on dry boards. (You also said c/r'ing is better on dry, so maybe you mis clicked) People c bet way more often on dry boards and it's much tougher for you to be balanced when you donk cause there's no draws to bet/ 3 bet. You're often WA/WB as well.