Coach Max, Poker Skill’s in-game coach, grades every decision in Practice vs AI as Optimal, Okay, Suboptimal or Poor. On this page a mistake is a decision he graded Suboptimal or Poor, the two grades Poker Skill’s report card counts as mistakes. From October 4 to 31, 2026, Coach Max graded 359,268 decisions by 5,219 Poker Skill players, and 18.7% of them were mistakes. The lesson and puzzle numbers further down come from the same four weeks. How we counted.
Which street do players get wrong most?
25.1%
The river. Coach Max graded 25.1% of Poker Skill players’ river decisions as mistakes, the highest share of any street, against 15.3% on the flop.
11,753 of 46,812 river decisions · 4,381 players · Practice games and Sit & Gos · Oct 4–31, 2026
Practice games deal some starting hands far more often than others, which can skew these shares. Sit & Gos deal the cards as they come, and there too the river had the highest share of mistakes and the flop the lowest: Coach Max graded 24.5% of river decisions as mistakes (4,699 of 19,210, by 1,094 players, October 17 to 31), against 13.6% on the flop.
By number rather than share, Poker Skill players made the most mistakes before the flop, the street where they made the most decisions: 28,090 of their 67,159 mistakes (41.8%), against 11,753 on the river.
Bets were Poker Skill players’ most common mistake on the river: 4,507 of the river’s 11,753 mistakes. Coach Max graded 37.6% of river bets as mistakes (4,507 of 11,989, by 2,474 players). In 83% of those mistaken river bets, the move Coach Max plays most in that spot was a check; the other 17% were the right move at the wrong size.
Why it’s a mistake. With no cards left to come, a river bet can’t protect a hand. It has to make money from the action it gets now: a call from a worse hand, or a fold from a better one. Checking is often better with a hand that can win at showdown but isn’t strong enough to bet: when worse hands won’t call, a bet gets action mainly from the hands ahead of it. The river is also the last betting round and the pot is usually at its biggest by then, so a mistake there can cost more.
Open the lesson in Poker Skill: Basic Betting Strategy
What is the most common mistake before the flop?
52.7%
Calling. 52.7% of the mistakes Poker Skill players made before the flop were calls, more than folds, raises and checks put together.
14,816 of 28,090 preflop mistakes · 3,625 players · Practice games and Sit & Gos · Oct 4–31, 2026
Coach Max graded 23.4% of preflop calls as mistakes (14,816 of 63,373 calls, by 4,874 players), against 11.0% of raises. In Sit & Gos alone, which deal the cards as they come, calls were 58% of preflop mistakes (7,946 of 13,713, by 1,023 players, October 17 to 31).
About 7 in 10 of the mistaken calls before the flop (10,823 of 14,816) came in spots where folding is the move Coach Max plays most, and the rest where raising is. About two thirds of the mistaken preflop calls (9,582 of 14,816) were limps: calls into a pot nobody had raised.
Why it’s a mistake. A raise can win the blinds on the spot, and a call can’t. Limping in puts no pressure on the blinds, lets more players see the flop and leaves someone else leading the betting. A call isn’t always wrong: calling a raise can be right, and when the action folds to the small blind, some strategies have it just call. But for most of these calls, folding was the move Coach Max plays most.
Open the lesson in Poker Skill: Raise or Fold (No Limping)
How often do players fold hands they should play heads-up?
31.5%
About 3 in 10. In heads-up Sit & Gos, Coach Max graded 31.5% of Poker Skill players’ folds before the flop as mistakes.
1,572 of 4,990 preflop folds · 685 players · heads-up Sit & Gos · Oct 17–31, 2026
Coach Max graded every one of the 1,572 mistaken folds Suboptimal, a small mistake, and none Poor. About 3 in 4 of the mistaken heads-up folds (1,177 of 1,572) were the small blind folding before anyone had raised, in spots where the move Coach Max plays most is a call or a raise.
Why it’s a mistake. Heads-up, the blinds are money already in the pot, and a raise can take it when the other player folds. So the small blind should play a broad range of hands rather than fold often.
Play a heads-up hand against the AI in Poker Skill
Which poker ideas are hardest to learn?
33.7%
Playing the board, just ahead of the rule of 4 and 2. In Poker Skill’s lessons, 33.7% of first answers to questions on playing the board were incorrect, about twice the 16.4% across all lesson questions.
2,994 of 8,887 first answers · 1,092 learners' devices · Poker Skill's lessons · Oct 4–31, 2026
On hand rankings questions, 26.4% of first answers were incorrect (2,647 of 10,033, on 1,637 learners’ devices), against 16.4% across all lesson questions. These numbers compare ideas inside Poker Skill’s lessons: how often an idea is missed depends on how a lesson asks about it.
The rule behind it. Each player makes the best five-card hand from their two hole cards and the five on the board, using both, one or neither of their own cards. When the board alone is the best five-card hand for everyone still in, they all play the board and split the pot. A kicker in your hand that isn’t one of those five cards doesn’t count.
Open the lesson in Poker Skill: Play the Board - Board Is Best
How often do players get poker puzzles right?
82.0%
Poker Skill players got 82.0% of puzzles right on the first try, with an Optimal Move or a Good Move.
27,841 of 33,942 first tries · 4,468 players · Oct 4–31, 2026
How often a puzzle is played right depends on how hard it is, so these numbers describe Poker Skill’s puzzles as well as its players.
Daily Puzzles were a little harder, counting earlier days’ puzzles as well as each day’s: Poker Skill players got 78.8% of them right on the first try (13,848 of 17,584, by 3,861 players).
Play today’s Daily Puzzle in Poker Skill
Where do you go wrong?
Your report card in Poker Skill answers that question for you. It reads your lessons, puzzles and hands in Practice vs AI and names your biggest leak. It also sorts your mistakes by which way they went: Calls too often, Folds too often, Bets and raises too often, Bets and raises too rarely and Wrong bet size.
How we counted
- Who. Poker Skill players. Every number is a total across players; none describes one player. The accounts and devices of Poker Skill’s team and testers are left out.
- When. October 4 to 31, 2026, Central time. The Sit & Go numbers cover October 17 to 31.
- Decisions. Each decision a player made in Practice vs AI, in Practice games and Sit & Gos, that Coach Max graded, counted once. A mistake is a decision graded Suboptimal or Poor; Optimal and Okay are not mistakes. Every grade compares a decision with Coach Max’s own strategy, in hands against the app’s AI players, so these numbers measure decisions, not how players’ hands turned out.
- Lessons. A question’s first answer each time a learner played its lesson, sorted by the idea the question teaches. Lesson numbers count learners’ devices.
- Puzzles. Each signed-in player’s first finished try at each puzzle, leaving out the puzzles in the app’s first-time setup. Right means an Optimal Move or a Good Move.
- How many. Every share here rests on at least 1,500 decisions, answers or tries, from at least 300 players or devices.
Citing these numbers
Quote a number with its dates and Poker Skill’s name, and link this page. A sentence you can use as it is:
Poker Skill players were likeliest to go wrong on the river: Coach Max, Poker Skill’s in-game coach, graded 25.1% of 46,812 river decisions as mistakes from October 4 to 31, 2026, against 15.3% on the flop (Poker Skill, “Where Poker Players Go Wrong”, https://www.pokerskill.com/blog/where-poker-players-go-wrong/).