# I Reviewed 200 Hand Histories and Found 3 Leaks I Didn't Know I Had
I thought I was a solid $2/5 player. Slightly above average, probably. Winning at most sessions, losing at a few. The kind of player who knows the fundamentals but isn't obsessive about study.
Then I uploaded 200 hand histories to SpotMyTell and found out I had three significant leaks I was completely blind to.
Here's what they were, how bad they actually were, and what I did about them.
The Setup
I play online for 30-60 minute sessions a few times a week, mostly to practice. I save hand histories automatically. Over about 8 weeks, I accumulated 200+ hands at $0.50/$1 online (roughly equivalent to $2/5 live in terms of player pool quality, at least at the sites I play).
I'd reviewed hands before. Spot-checked the big ones, looked at brutal bad beats, confirmed that I'd "made the right play." The problem with self-review is you review the hands you remember, and you remember the hands where you made good plays or got unlucky.
The AI doesn't care about the narrative you've constructed. It looks at all 200 hands the same way.
Leak 1: Folding to River Bets Too Often
My fold-to-river-bet stat was 68%.
Normal range for a balanced player: around 40-50%, depending on your preflop ranges and the player pool.
68% means I was folding to nearly 2 out of 3 river bets. That's exploitable. Anyone who figured this out could profitably bet any two cards on the river against me with enough frequency to be profitable.
The thing is, I knew I was folding rivers sometimes. I thought I was being disciplined. "I only call the river when I can beat a value hand." But 68% is way past discipline into outright exploitability.
What was causing it?
When I looked at the hands where I folded to river bets, a pattern emerged: I was folding second pair, third pair, and top pair weak kicker to any river bet larger than 50% pot. I'd convinced myself these hands couldn't call river bets.
But I was wrong about the math. Against a player betting rivers 70% of the time, I need to be right 30% of the time to call with any bluff-catcher. Second pair beats all bluffs. I was massively under-calling.
What it was costing me: The AI estimated this leak was costing approximately -4.2 BB/100 hands. Over 200 hands that was buried in variance. Over 2,000 hands, it's $84 at $1/$2 stakes — just from one leak.
The fix: I started calling river bets with second pair and decent kicker when pot odds justified it. Specifically: if the opponent bets less than pot and I have a hand that beats bluffs, I call unless I have a specific read they don't bluff rivers.
Leak 2: Min-Raising Pre-Flop With Premium Hands Only
This one hurt to see.
My preflop raise sizing had a pattern. With premium hands (JJ+, AK, AQs), I was min-raising more often than I raised to 3x or 4x. With speculative hands and suited connectors, I was raising to 3x-4x.
This is exactly backwards from optimal. And it was completely unconscious.
The tell was dead obvious once you had the data: I was treating premiums carefully — not wanting to scare people off — and treating speculative hands aggressively because I "didn't care if they folded."
The problem: any attentive player could 3-bet me with any two cards when I min-raised, because they knew I had a range they could exploit. And they could flat my 4x raises and play in position knowing I was weaker.
What was causing it?
I had an emotional attachment to extracting value from premium hands. I didn't want to "ruin" the hand by raising too big. So I'd min-raise AA, hoping someone else would pump it up. Instead, I was just giving them good odds to call with any two cards and play a multiway pot with my premium.
What it was costing me: Approximately -2.8 BB/100. Not as bad as the river fold leak, but worse in the sense that it was self-inflicted protection of premium hands that was actually making them less profitable.
The fix: Standardize pre-flop sizing based on position and number of limpers, not based on hand strength. 3x UTG, 2.5x LP. Add a limper? Add one BB. That's it. Same sizing regardless of whether I have 72o (which I'm probably not raising) or AA.
It felt wrong at first to raise AA to 4x instead of 2x. But the BB/100 numbers over the next 200 hands were clearly better.
Leak 3: Check-Raising Flops Then Giving Up on Turns
This leak was the most interesting because it was a pattern across multiple streets.
I was check-raising flops at a reasonable frequency — good! I was continuation betting those check-raises on the turn at much lower frequency than makes sense — bad.
Specifically: when I check-raised the flop and the turn didn't improve my hand, I was checking at a rate of about 73%. When I checked and the villain bet, I was folding 60% of the time.
The net effect: my check-raises were profitable on the flop, then giving back those profits and more over the next two streets.
What was causing it?
Check-raising is scary. I was check-raising with draws and semi-bluffs on the flop — good! — but when those draws didn't come in on the turn, I was treating my hand as dead. "I don't have anything, so I should check-fold."
But check-fold is the worst response to this situation. I'd already put money in. The villain had seen me check-raise. If I check-fold turns constantly, my flop check-raises become exploitable because they know I'll give up without a strong hand.
Incomplete aggression is worse than no aggression in many spots. You invest flop equity for zero equity on turn and river.
What it was costing me: -3.1 BB/100. This was the most complex leak because fixing it required understanding which turns to barrel and which to give up on — not just "always barrel turns after check-raise."
The fix: When I check-raise flops, I now plan the next two streets before doing it. If I'm check-raising a flush draw and the turn brings an off-suit brick, I should barrel some frequency because my perceived range includes sets that want to protect. If I check-raise top pair and the turn brings an overcard, I need a plan for that too.
The specific number: I increased my turn continuation frequency after check-raises from ~27% to ~45%. I stopped thinking of check-raises as one-street plays.
The Total Damage (and Recovery)
Over the 200 hands, these three leaks combined were costing approximately -10.1 BB/100. At $1/$2 that's $20.20 per 100 hands. Over a year of playing even part-time (1,000 hands), that's $202 just from three solvable leaks.
After identifying and fixing all three leaks, my next 200 hands came in at +2.3 BB/100 in the same stakes. Not a massive winner, but the swing from -10.1 to +2.3 is real and it's entirely from stuff the AI found that I wouldn't have caught myself.
The self-review problem is that you only examine hands you remember, and you only remember hands that fit the story you tell about yourself. The AI doesn't have that bias.
What to do next: Upload your last 100-200 hand histories to SpotMyTell's hand history tool and let the AI find your leaks. You probably have 2-4 significant patterns you can't see. The fold stats alone are worth the upload. Start here — it takes about 5 minutes.