Half your ICM flop folds are just nitty - Greg Kozieja

COACHING RECAP - GREG KOZIEJA

A lot of the 'ICM' folds you make on the flop aren't really about ICM at all. You're just playing scared. That was Greg Kozieja's blunt take in his June 3 session on flop check-raising with risk premium - and it starts with knowing when ICM even matters.

First, what this is even about

The whole session lives in two words - risk premium - so it's worth slowing down on. In a cash game, a chip is a chip: lose your stack and you reload. In a tournament you can't. Bust out and you're gone. That means losing your chips hurts more than winning the same amount helps you, and that gap is the risk premium. It grows the closer you get to the money.

Greg picked the topic because it's finally something you can study properly. GTO Wizard's AI can now solve these post-flop spots, and fellow coaches Fabi and Curtis had already started digging in. His twist was to change one thing at a time - keep the hand identical and only turn the risk premium up - so the lessons carry into any spot you actually sit in.

Most players reach for ICM far too early

Here's the pattern he keeps seeing. You're deep in a tournament, a third of the field still to bust, you flop a so-so hand and fold it 'because ICM'. It feels disciplined. But when Greg runs the numbers at that stage, the right play barely moves from normal poker.

You overfold on the flop ICM 50 and justify it as ICM impact. You're just basically saying that I'm a nit.

It only really starts to bite once about a third of the players are left, and it reshapes everything by the time you reach the bubble. Before that, you're mostly just folding too much.

Worth hearing him out on this: Greg has $1.26M in live earnings, a EUR 474k Eureka title, and made his Triton debut at Jeju 2026 - after coming up through the Pokercode community and Grindhouse 3 to become a coach himself.

One spot: the bubble changes which hands you raise

The whole session runs on one fixed setup: 25 big blinds deep, button against big blind, equal stacks. Greg keeps that frozen and only turns the risk premium up, so you can see what the bubble alone does to a hand.

Take the flop Ace-King-six, played from the big blind. For chips, you check-raise a wide range here - two pair and better, but also a top pair like ace-seven, and even some sixes as bluffs.

Move the same spot to the bubble and the top-pair check-raises almost vanish. In Greg's 15% risk-premium solve, the only top pairs still raising are ace-nine and ace-eight of hearts - the highest-equity ones. Most of what used to raise now just calls.

But 'just never raise' is the real trap

The tempting next step is to raise nothing at all. Greg says that's the mistake: you'll still flop two pair, and those hands want to raise and get chips in. Cut your raises to zero and your range becomes easy to read.

playing zero check-raise is not the answer. It's actually something that you wanna avoid.

What else is in the session

From there it gets fun. Greg runs the same idea across three more flops - King-six-five, Ten-seven-three and a tricky paired six-four-four - shows why your bets quietly shrink to min-bets as the money nears, how to spot the one flush draw actually worth bluffing, and why he'd ignore GTO Wizard's earliest ICM model almost entirely.

This is the first in Greg's risk-premium post-flop series, with probe-bet lines coming next month and companion sessions from Curtis and Fabi already on the platform - all inside the Pokercode membership, with subtitles in DE, EN, FR, ES and PT.

The same coaching that put Greg on the Triton stage. Join free first - no credit card required. pokercode.com