Interesting cbet study

I open a ton of hands on the btn and co (this is not a thread about TAG v LAG) and decided that by shrinking my opening raises to 2.25bb and 2.75bb from btn and co respectively, that I'd make some cheaper steals. It did not work. The steals stopped working and my cbets didn't work at all in these smaller pots.

The smaller opening raises began the second week of june just after the peak and I continued to use those raise sizes until this week. It could just be normal variance but I think theres something to this. Maybe fish get scared in slightly larger pots or they assume that I have nothing if I'm stealing with a tiny raise and will get sticky with a lot more. Anyways, I'm experimenting with 3.5bb preflop raises now from all positions and the cbet has suddenly started working again.

I probably should have put the graph in BB's but I play 4nl so this represents a lot of bb's.

*this is just a cbet results graph

Attached Images

cbet graph june.jpg (17.5 KB, 57 views)

Personally I open 3x plus 1 per limper,

Cbet at micros, i bet about 60% when i miss and 70%+ when i hit dependant on board texture.

I think sample size + variance.

Its fine opening 3x, 3.5x, 4x from all positions at 4nl, and my default would be 4x from all positions and 3x on the button at any limit below 25nl probably ,just because games are not aggressive. You should always plan your raise sizes on your opponents stack sizes and tendencies though.

If you really think people are folding less to c-bets in these spots, barrel more... simple! Really though, if anything people will just call a bit more with speculative hands, you're still gonna get a ton of folds from people calling in the blinds if you're c-betting on the right boards.

This is may's cbet graph. For the most part was doing 3bb btn open, 3.5 co, 4bb utg+1, and 5bb in ep. I really never had a down swing. The only downswing my cbets have ever had(of sufficient duration) occurred when I started opening smaller. Yes, I think they were calling in the blinds with speculative hands, but I think post flop they were getting extra sticky with their bottom pairs and draws because they refuse to believe I'd open AK for 2.25bb or some fishy rationaliztion. Together these show my cbet results over the past 40k total hands.

Attached Images

cbet graph may.jpg (17.4 KB, 34 views)

I don't think fish respond to cbets the same way that thinking players do. I don't think they look at a board and think, "if I'm behind then I'm way behind, should fold/bet fold/check" Its more like, "ace high flop, hmm, no way he has an ace this time cuz he had it last time, my sixes are probably good. OOo board doubled, now I have two pair defo have him beat." Stuff like that. They think about what we could have not what we're likely to have(just my theory thus far). Its still a smal sample size, but over the 8k hands(like 3.5k I used smaller raises) my overall game was breakeven(which is a good result when experimenting) but its the first time my cbets stopped working in 40k hands. Also my cbet success rate during the period of small raises was 38%(i experimented with different cbet sizing as well but i think its negligible) and with big pf raises it is 44-46%. And I was actually cbetting less during the small raise period(68% during small 73% otherwise). Still could just be variance but someone else should do a confirmation experiment.

So when your bet size pre was bigger you were getting more folds pre and post when you cbet? Maybe they put you on a stronger range (hand!)? Whereas when you made it smaller they thought their bottom pair was probs good? Dunno, but it could be as simple as that no? Also, as acky says, the sample is small.

I have been experimenting with funky betsizing vs regs, by accident and on purpose, and the results have been good so far. Let's all do things differently in order to beat the standard bet sizing/lines! Wooooooo!

Yes, exactly. And yes, I think its probably that they put me on a stronger range.

Honestly, Sample size is so small you really cant make comparisons. I really dont think fish are thinking enough to start folding significantly more when theres 1bb more in the total pot on the flop. 40k hands is a lolsample when you include both raising smaller and raising larger and you say you only tried raising smaller for 3.5k hands? It would still be meaningless if you raised smaller for 35k hands, the only way to get something meaningful (yet probably still really insignificant because its so unimportant and there is not going to be a huge difference between the two strategies) would be to try both for the very minimum of 100k hands. People fold to c-bets based on their hand and how it connects to the board, they arent thinking oh he raised 2.2x on the btn and its Axx i will not fold 66, because even if you raise 3x on the btn and its Axx they still wont fold 66 because they have second pair or whatever, and they very well might be outplaying you by calling with 66 on these flops if you arent going to be firing multi street bluffs because the chance you have an A when you raise on the btn is quite low if you're stealing a decent percentage.

Basically i just think you're over-thinking this. People are bad, raise 3x at micro's to make more money in position because people hardly 3bet. Raise 2.5 or 2x if two shorties in the blinds that 3bet like 15% each. Otherwise don't worry about it, steals should work less when you raise smaller, but they dont have to work as much either, c-bets are going to work just as much with either over a good sample imo.

Also, i'm not sure what the point in raising 5x UTG, 4x, in HJ, 3.5x CO, and 3x in the btn is. Yeah its good raising bigger with tighter ranges but 5x just seems a bit big and you're ranges arent going to change so significantly between positions, especially UTG and the HJ (i guess you're playing 6max) I guess what i dont like the most is the 5x UTG. Keep it simple

lolsample. Thats funny.

But yah I totally get it. It was a sample of 3k cbets tho not hands which yes, is still really small, but I hadn't had a period of losing cbets longer than 250 cbets over the past 40k hands and then suddenly I had a cbet downswing that lasted 3.5k cbets, so considering the limited data, it seemed like a statistically significant change that correlated exactly with when I changed my raise sizes. Thats why I started thinking about what the hell changed and the raise size seems like the most(and pretty much only) significant change in how I was playing. Its a correlation thing because I know I can never really prove causation. And fish are all different so even if one thinks in a predictably stupid way, its unlikely they all are. I'll come back when I have a bigger sample. Its more than likely that I'm over thinking it(better than not thinking about it) and maybe its even worth thinking about.

And the escalating raise sizing from by position in May was just another experiment based on the idea of making their odds worse to call me when they'll have position.

Btw, what kind of swings do you guys have in your cbet results? Obviously at higher stakes you'll have bigger swings but my cbet graph seems amazingly low variance except for my baby raise experiment which is still pretty consistent just consistently negative.

Plus I'm hoping that after 100k hands I'll be at another stake where I'll have a whole new batch of experiments.

hehe, FWIW, you should probably make it atleast 3x on the button when 70bb+ fish are in the blinds anyways, because c-betting is so profitable and so is value betting. The whole lower your raise size is only really relevent to 40bb fish (when you should minraise or 2.5x as there is no need in 3x'ing with any hands because you can still get stacks in) and when we're against regulars who we want to steal against a lot who arent just going to let us run them over and shortstackers who 3bet a ton. So basically everything in your little study is irrelevent about the psychology of fish folding to cbets in 8c smaller pots because you should probably be raising bigger to make more moneys vs these guys ?

I have defo not been adjusting my raises to stack sizes. Will start taking that into account.