Category: Uncategorized

Ten-three of clubs, from the big blind, for a world championship. That was the last hand of the 2025 WSOP Main Event, and if you want a single image for how Michael Mizrachi played the deepest stack of his life, that’s it: a hand nobody defends out of habit, played in a spot where the price and the position made it correct anyway.

The run itself was absurd. Mizrachi beat John Wasnock heads-up out of a 9,735-entry field for $10 million and his eighth bracelet, in the same calendar year he won his fourth Poker Players Championship. The final table lasted 79 hands. He started the final day with roughly three-quarters of the chips and busted two players in the first two hands.

But look at how he got the stack, because that’s where the lesson hides. The pivot was ace-king against Wasnock’s pocket kings — a cooler he won by spiking an ace on the river, flipping 76% of the chips in play to his side. Deep-stack poker at a Main Event final table isn’t a sequence of clean edges. It’s a handful of enormous, variance-soaked decisions surrounded by dozens of small ones, and the small ones are what buy you a seat at the big ones.

That’s the actual takeaway on hand selection. Mizrachi wasn’t playing wider because he had chips; he was playing wider in the specific spots where a big stack converts range advantage into real money — big blind defends at good prices, button raises against short stacks who can’t fight back, and near-zero interest in marginal flips when he was already winning. Wasnock, to his credit, fought back, taking three of five pots at one stretch and even sniffing out a Mizrachi bluff. Then he raised the button with ace-nine, flopped top two, and ran into a big-blind defend that turned a flush.

Deep stacks reward players who widen where it’s cheap and tighten where it’s expensive. The Grinder’s ten-three wasn’t a hero call on intuition — it was a hand he was always going to see a flop with. Same principle as protecting a monster stack without lighting it on fire: the chips buy you leverage, not license.

Photo by Tima Miroshnichenko on Pexels

The chart isn’t wrong. That’s the annoying part. When your app tells you the button opens J4s, 86s, and 64s, it is telling you the truth about a very specific universe — one where everyone folded, stacks are dead even, you’re opening 2.5bb, and the guy in the big blind defends and then plays a competent turn. Walk into a $1/$3 game on a Friday night and count how many of those conditions actually hold.

Usually: zero.

There are three limps. Someone straddled. The blinds are 150bb deep because they bought in twice. And your 2.5bb open now has to squeeze through four players who will call it because they’ve already got money in and they like the look of things. That’s not the tree the solver solved. A long-running thread on the live low-stakes forum puts it plainly — the lower stakes are more beatable with exploitative play, and the same discussion flags the real mechanism: equilibrium strategies earn the most against opponents who are also near equilibrium. Against a table that isn’t, you’re leaving the surplus on the felt.

The thread also surfaces a point I keep coming back to — bots playing pure equilibrium still win at low stakes, but sometimes not by enough to clear the rake. That’s the whole ballgame in a live room. Your GTO-approved 64s open is a marginal EV hand in theory and a negative one after a $6 drop plus jackpot, played multiway, out of position, by a human who isn’t a solver.

So what survives? Ranges that punish the actual mistakes in front of you. Tighter and more positionally honest when there are limpers, because isolating three passive callers with a hand that needs to flop perfectly is a donation. Much wider on value — your big pairs and strong broadways print more here than any chart accounts for, because the calls come from ranges that never fold. Bigger sizings, because nobody is price-sensitive.

If you want the postflop half of this problem, my notes on multiway solver ranges in 3-way pots cover where the tools are finally catching up.

The chart is a map. The room is the territory. Read the room.

Photo by dp singh Bhullar on Pexels

Somewhere between the pandemic reopening and now, $7 out of the middle became $8, and nobody sent a memo. Players in Vegas have been reporting that Horseshoe Las Vegas moved to a $6+$2 structure, with similar moves reported at MGM properties — the kind of change that reads as a rounding error and plays like a tax hike. Depending on whether you count the promo drop, that’s somewhere between a 14% and 20% increase.

The mechanics matter here. A “$5+$2” rake means a $5 cap on the house rake plus a separate $2 promotional drop, usually taken once the pot reaches $10, with the normal rake running 10% — a dollar per $10 in the pot until the cap. That means at $1/$2, almost every pot that sees a flop gets fully raked. The cap isn’t a ceiling you rarely touch; it’s the default.

Run the arithmetic. Thirty hands an hour, call it $6 average dropped per hand once you account for the walks and folded blinds, and your table is bleeding roughly $180 an hour into the box. Split nine ways: about $20 per player, per hour. Add dealer tokes — three pots won an hour at a buck each — and you’re at $23. In big blinds, that’s a hair under 12bb/hour of pure friction before a single card is dealt.

So a “good” live reg winning 8bb/hour net is actually beating the table for 20bb/hour gross. That’s an enormous edge, and it exists only because $1/$2 lineups are genuinely bad. The second that stops being true — the second the table has three competent regs instead of one — the game is unbeatable at $8 a pot.

Bankroll accordingly. Live variance sits around 90bb/100 hands, and a thinner post-rake win rate stretches your downswings. Thirty buy-ins used to be a fine number for $300-cap $1/$2. In 2026, budget forty: $12,000. Not because you’re worse, but because your EV per hand is smaller and the noise hasn’t shrunk with it.

Two adjustments actually move the needle: play more pots in position where you can win uncontested, and stop flatting garbage multiway — the multiway math is uglier than most people assume. If you want a live-specific framework, the one laid out in Ed Miller’s The Course still holds up, and SplitSuit’s breakdown of live cash realities is a useful sanity check on what rooms are actually charging.

Photo by Burst on Pexels

Big stacks kill more players than short stacks do. Not literally — the short stack busts — but the monster stack is where tournament equity goes to die quietly, one “I have chips, I can afford it” call at a time. Which is why Lucas Jumalon’s 2026 WSOP Main Event win is worth a slow re-watch.

He went wire to wire. Nine-handed, 328 million in chips at the start of heads-up play against Lauri Saaskilahti’s 225 million, $10 million and a bracelet at the end of it, out of a 9,208-entry field. Second-youngest champion ever, behind Joe Cada. All of that is the headline. The lesson is in the texture.

First: he actually used the stack. The Review-Journal noted that he tried to bully the other eight players with his big chip stack on the first day of the final table, and it worked — not because bullying is clever, but because ICM at a Main Event final table is brutally asymmetric. Every pay jump is life-changing for the guy with 15 blinds and rounding error for the guy with 100. The chip leader’s edge isn’t that he can gamble; it’s that nobody else can. Jumalon pressed exactly that seam, particularly in the limped and multiway spots where short stacks are desperate to see a cheap flop and terrified to commit.

Second: he was willing to be wrong without being reckless. On hand 166 he moved all in with A-9 into Saaskilahti’s A-10 and watched a runner flush end the hand, the closest he came to surrendering the lead all final table. That’s the cost of an aggressive game plan, and the tell of a good one is that losing that pot didn’t change the plan.

Third — and this is the part amateurs skip — he still made people pay to find out. The championship hand had $100 million already in the middle on 4♣A♠K♠J♠6♠, Jumalon holding the Q♠ for a flush, and he shoved 180 million rather than milking it. Saaskilahti hero-called with J♦3♦.

The monster stack isn’t a cushion. It’s leverage with an expiration date. Use it while the ICM pressure is pointed at everyone else, keep your value bets big, and accept that occasionally the river burns you. That’s the whole curriculum.

Photo by Jonathan Borba on Pexels

Multiway Solvers Are Here: What GTO Wizard’s New Preflop Tool Actually Changes About Your 3-Way Pots

For years, the honest answer to “what’s GTO when three of us see a flop?” was a shrug and a heuristic. That era is quietly ending. GTO Wizard has launched custom Multiway Preflop Solving for up to nine players, built on top of its earlier heads-up and 3-way postflop tools, and the pitch is that instead of waiting hours on desktop hardware, you solve multiway preflop spots in your browser — or on your phone — in seconds. The engine behind it combines traditional CFR solving with neural-network value estimation rather than leaning on static precomputed blueprints.

Why this took so long is worth understanding. Adding a third player can inflate computational complexity by a factor of a thousand or more. And there’s a philosophical wrinkle: unlike two-player chip-EV spots, optimal play multiway is an ill-defined problem — a Nash equilibrium exists, but it doesn’t carry the same guarantees as in two-player zero-sum games. Players can gang up on each other, intentionally or not, and if two opponents drift to a different equilibrium, the third player is the one left holding a worse result. Treat multiway output as a strong map, not gospel.

So what actually changes at the table? The most useful corrective is directional. Three-handed play tends to run more passive than heads-up baselines suggest: with two players left to act, the cost of aggression rises and bluff frequencies drop meaningfully across streets. A board like J♠T♠5♥ that begs for barrels heads-up sees far more checking three-way, and players importing heads-up sizing habits are simply overbluffing. Squeezing flips the other direction — bigger size, tighter and more polarized, because both opponents have to fold for it to work.

The implementation details matter too. The 3-way engine supports asymmetric stacks, side pots with all-in players, any rake structure, and the bunching effect — the things live and MTT players actually deal with. ICM and bounties are still on the roadmap.

Bottom line: your multiway instincts were probably too aggressive postflop and too timid with your squeezes. Now you can check.

Photo by Pavel Danilyuk on Pexels