Super Bowl Sunday Special #5 Our First Cash Game Hand
David Steinberg makes a flush blind vs blind.
Today’s punter is David Steinberg is a micro-stakes poker player and a Professor of International Relations at Johns Hopkins University. You can learn more about his academic writing on his website. He hopes, in the future, to combine his two passions and write a book about what social science and poker strategy can learn from each other.
My thoughts are included in the footnotes. If you reading this via e-mail, it might be an easier read on Substack where the footnotes require less scrolling back and forth. Click here.
If you are interested in being the subject of a future Sunday Special, let me know. Do not be shy if you have a lack poker skill or accomplishments. No solver analysis is required from you and I’d much rather have amateur poker players, who are good writers that can produce clean copies and clearly articulate their thought process than editing the writing of 99% of accomplished poker players. Onto David’s analysis.
As a “serious” recreational playing micro-stakes poker, to say that there are some gaps in my poker knowledge would be an under-statement. Like many recs that have only limited time to study, I know my preflop ranges pretty well, and the most common spots like c-betting in position in single-raised pots. I am also competent enough to not mess up too badly in relatively easy post-flop situations, such as when I have the nuts or when I have obvious bluffing candidates like the nut flush blocker. But, like most recreational players, I often struggle playing out of position in wide configurations such as blind versus blind. Picking the right check-raise bluff candidates when many are available can also be tricky, as is playing hands that are strong but far from the nuts. The hand below touches on each of challenges. Hopefully you can benefit from my mistakes and avoid these types of blunders yourself.
10 NL Cash Game on Global Poker, 6 handed
LJ, HJ, CO, and BN fold.
Hero has $13.69 in SB and BB player covers with $16.40.
Hero has 4s3s and opens to 3 big blinds. Big blind player calls.
***FLOP*** [Ts, 2s, 9c]
Pot size: $0.60
Hero: check
BB: Bet 0.18
Hero: Raise to 0.86
BB: Call 0.68
***TURN*** [Ts, 2s, 9c] [5s]
Pot size: $2.32
Hero: Bet $1.74
BB: Call
***RIVER*** [Ts, 2s, 9c, 5s] [6h]
Pot size: $5.80
Hero: Bet $4.35
BB: Call
BB shows 8s6s and collects $13.78
What I thought during the hand
Preflop:
I have looked at enough preflop charts to know that 43s is not a GTO open. But I was trying to open wider from late position. I didn’t think this hand would lose much EV in theory and I figured at the time that my postflop edge over villain should make this at least breakeven, if not slightly profitable, to open.1
Flop:
I decided to start with a check. I didn’t think I had a massive range or nut advantage on this board. I also know that I want to check a lot of my range when out of position.2
Facing a small bet, I deliberated between call and raise. I decided to raise by leaning on a heuristic I had acquired, that hands that have no showdown value and high equity make for good (check-raise) bluffs.3
This hand is literally the nut low, with zero showdown value at this moment. I believed that my flush draw and backdoor straight draw meant that the hand had good equity. So, I decided to raise, clicking the 75% pot raise size, admittedly without a solid grasp on optimal check-raise sizings.4
Turn:
I make my flush, and also pick up an open-ended straight flush draw to boot. I decided to barrel, betting 75% pot. My thinking in game was that I can target many of villain’s worse hands with this sizing, like top pairs or two pairs. I also figured that the turn card was good for my range overall since it completed one of the obvious draws that I would be check-raising.5 Since I polarized already on flop, this somewhat polarizing bet seemed appropriate. However, I also recognized that flush-completing turns tend to reduce bet sizing overall, so I did not think an overbet was appropriate for my range.6
River:
The 6h doesn’t change anything (other than completing my backdoor straight draw, which doesn’t actually improve my hand’s value). I bet 75% of the pot again, with the goal of getting called by hands like AT, T9, or 22. I considered jamming for around 2x pot, but didn’t think I could get called by enough worse hands, and this size would just isolate myself against better flushes. I thought a 75% pot would allow me to still win more than half of the time that I was called, making this a profitable value bet.
Post-Game Analysis
One challenge for reviewing this hand after the fact is that 43s is not in any standard solver’s SB range. To review the hand, I first looked up how GTO Wizard plays similar hands such as 64s.7 To complement this, I ran my own custom sim on free solver software, where I inputted my own wider SB opening range as well a more typical micro-stakes player’s BB defense range. (For the latter, I include hands like Q8o, T5s, and 76s at full frequency since this population rarely three bets these hands). My custom solve should be taken with a giant grain of salt, but both analyses generate nearly identical patterns, and none of the conclusions that follow are based solely on my own sim.
Preflop:
My hand, 43s, has an EV of about -0.02, meaning it loses 2 percent of a big blind on average. As expected, then, playing this hand does not lose much EV in equilibrium. If – and this is a very big if – I could play this hand well post-flop, it would be reasonable to include 43s in my SB opening range. But, in case it isn’t obvious already, my analysis of my play on subsequent streets casts serious doubts on this assumption.
Flop:
SB has a slight equity advantage on this flop, around 52%. Overall, SB checks around 75% of its range on this board. Low flush draw hands like mine mostly check, but they mix in some bets. This may be the one decision I did not get wrong. The fact that most hand combos mix between bets and checks sure helps me avoid an error on this node!
Check-Raise on Flop:
The very general idea that motivated my check-raise still seems reasonable. The decision to check-raise was motivated by an understanding that hands with low showdown value and high equity should bluff. My interpretation of the solver output is that this heuristic works out pretty well. Hands that are strong draws but have considerable showdown value like AsKs and AsQs rarely raise in this spot, presumably because they are too strong to turn into bluffs.8 The most frequent bluff hands are those like QsJs, Qs8s, Js8s, Ks8c, and Qs8c that have less showdown value but still have considerable equity.9
I did, however, make a large mistake in assessing my hand’s equity. My hand (4s3s) and others like it (5s4s, 6s4s) only have around 42% equity. It is hardly a surprise then that these holdings almost never check-raise the flop.10
My key take-away here: you are much better off semi-bluffing on the flop with a hand that has mediocre showdown value but good equity than with a hand that has absolutely no showdown value with mediocre equity. In other words, the solver does not advise turning the absolute nut low hand into a check-raise bluff on the flop, at least not in this case.11
Turn:
The flush-completing turn obviously improved my hand, and I thought it improved my range as well. The latter part turns out to be wrong. Both players have around 50% equity on most spade turns, including this one. Villain’s range that calls my flop check-raise contains many flush draws—a point that should have been more obvious to me in game but was not. In fact, in the GTO Wizard sim, flushes make up a larger portion of BB’s range than SB’s (20.5% vs 17.5%).12
Lacking a range advantage, SB actually checks the majority of range on this card. And, when SB bets, the main size is small, around one-third of the pot. Larger bets are almost never used, likely because neither player has a clear nut advantage here either. Additionally, my particular holding seems far too weak to bet big. My large turn bet is undoubtedly a blunder.13
River:
Since my play on previous streets took us far away from the equilibrium path, the solver does not seem terribly useful in guiding river decisions after arriving here in this manner. But, my best interpretation of the solver output is that betting my flush for 75% pot on the river is reasonable. In the 100 big blind GTO Wizard sim, SB is betting all in for around 80% pot with all flushes, straights, and sets. If my 75% pot bet was a mistake, it was not the biggest one I made in this hand.14
Overall Assessment and Lessons
Since this hand occurred in a 10 NL game, this punt is small in terms of actual dollars. Yet, it cost me far more big blinds than was necessary. I took away three main lessons from this hand review that I hope will help me – and you – avoid throwing away our hard-earned pennies in the future. First, do not overvalue low flushes when, due to heavy filtering on earlier streets, your opponent’s range contains many stronger flushes.15 Second, when choosing hands to check-raise bluff, it is better to use hands with reasonable equity than to use the hands in your range with the absolute lowest showdown value.16 Third, if you are prone to making these types of errors, think twice before voluntarily entering pots with hands that have negative EV in equilibrium.17
I have not spent much time studying no ante preflop solutions, but I know in high rake games you often play raise-first-in or fold from the SB and I’d guess that 43s is a reasonable open here.
A good instinct from you, especially compared to most players who c-bet too often. Betting any flush draw, especially one with a three straight that is currently nut low is fine and might be best vs some players who might not find some loose floats.
This is an okay heuristic, but bad flush draws are a tricky class of hand to play aggressively. When you’re shallow you don’t want to check-raise-fold and you don’t want to check-raise- call a shove and stack off four high. When you’re deep stacked getting flush over flushed is a real concern, so you’d prefer checkraising a hand like Qs3s compared to 4s3s. Q3 can turn top pair and is ahead of hands like 8s7s. However taking an aggressive action with a flush draw can never be that big a mistake.
Generally speaking people check-raise to too small a size when deep. You are usually representing a very polar range that wants to get a lot of money in the pot. Pick a big size that actually puts stress on hands like top pair. 75% is a good size.
It’s good for your range in that it improves many of your bluffs, but it also decreases the EV of many of your value hands. Generally when flushes fill and you are deepstacked you play cagey and there is a lot of small betting on the turn, but a lot a lot of money can go into the pot on the river.
This is correct, but I still think 75% is a touch too big, but it’s fine to get greedy with a flush playing $10 NL.
43s does raise in the 100BB GG 1k NL Rake agent. So it gets in there depending on the rake and other factors.
Another factor here is that AsKs blocks all the big FDs the BB might have. As3s is more likely to get in a lot of money vs Ks4s than AsKs vs 4s3s.
Basically you want a draw that is strong enough to check-raise stack off or one that you don’t mind check-raise folding and 4s3s doesn’t fit either bill. You also like having an overcard to the ten.
The joke I’ve heard about these sorts of hands is they have 40% vs the nuts and 40% vs total air. So they have okay equity vs everything, but never have good equity.
Another problem here is if you miss your draw on the river, you don’t want to bluff the river with spades. So you often bet the turn with nut low, get called and then need to decided if it’s worth bluffing to get 87 or 6s5s to fold on the river, while knowing you block hands like As3s and As4s that will always fold.
This is another side effect of the big flop check-raise size. You start folding out some strong hands right away, which means cards like flush completers hit the BB’s range even harder.
I’d agree it’s a blunder, but in low stakes games betting too big with a good hand is generally a good strategy. However people in low stakes games also tend to bet too much middle of range when checked too, might not find overbet all-ins with nut flushes, etc. So checking still works
I’d agree with this
I think one other thing you’re missing here is just that after this line you’re representing a flush. He will have hands that can bluff catch like sets, two pairs, AsTx, etc. but you’re representing a flush and you have the worst one. It’s pretty hard to have more than 50% equity when called when in that situation
I think this is more of a flush draw dynamic than general draw dynamic. E.g, check-raising 54 on T32 is a high frequency play even though it’s nut low and has bad equity, but it has a draw is to the nuts. The problem here is that your draw is to a hand that is … in theory a strong bluff catcher or a dicey value bet if you end up playing for stacks. So it’s not just about the equity of your draw, but the EV of your hand when you hit your draw.
Neutral EV depending on rake effects. :)