tournaments
A tournament is not a long cash game.
A decision on 12 big blinds has nothing in common with the same spot on 100, the blinds grow whether you play or not, and the chips you win are not money until the tournament ends. Measuring tournaments with cash tools produces confident numbers that mean nothing. Everything below is separate on purpose.
four things kept apart
Two of these were repairs. The score ran on the cash formula, and the return on buy in ignored re entries, which flattered exactly the players who re enter most.
A separate score
Tournament hands never touch the cash score. They carry a phase, and that routes them to a tournament score with its own weighting: a base of 30 points, up to 60 for return on buy in, up to 40 for how often you finish in the money, up to 40 for push or fold accuracy, and up to 30 for open tournament leaks. Below thirty tournaments every one of those is pulled back toward the middle, because that is where the numbers stop meaning much.
Chips stay chips
A tournament hand is counted in chips, and the chips are never converted into euros. Chips are worth different amounts at different moments in a tournament, and without a full ICM calculation any conversion would be a guess dressed as an amount. Chips do act as a gate: a pattern that gained chips is not reported as a leak.
Cost is one buy in per entry
Return on buy in used to be computed from the buy in alone, which flattered anyone who re enters. Your stake is now the buy in multiplied by the number of entries, so a tournament you entered twice costs twice as much, both in the overall figure and per buy in level.
Phases are not invented
From a single hand only the blind level is reliable, so a hand is marked early or middle and nothing more. The bubble, the money and the final table depend on how many players are left against the payout structure, and that only lives in the result file. Reading them off a blind level would be a guess, so it is not done.
Both the hand histories and the result summaries come out of the same export, which is read as described on the import page, and the score they feed is explained component by component on the Poker IQ page.
the tournament detectors
Four families with their own minimum counts. Where the chips can be traced to the hands in question, a losing balance is required as well: a pattern that made chips is not reported, however unusual its frequency looks.
Calling without raising
- · At least 60 hands played voluntarily outside the blinds
- · More than 15 percent of them a call with no raise
- · And those hands lost chips on balance
In a tournament every chip you put in without a plan shortens your life, and a caller cannot win the pot immediately.
Calling on a short stack
- · At least 25 voluntary hands on 12 big blinds or less
- · More than 20 percent of them a call rather than a raise or a shove
- · And those hands lost chips on balance
On a short stack a call has the drawbacks of both other options: you spend chips without being able to win the pot and keep too few to apply pressure.
Too passive after the flop
- · At least 60 tournament hands that reached a flop
- · You bet or raised in fewer than 40 percent of them
- · Frequency only, with no money condition
The loss here lives in pots you never built, so it cannot honestly be traced to individual hands and carries no price at all.
Raising too little for the hands you play
- · Measured per stack depth, in four brackets from under 15 big blinds upward
- · At least 150 hands in that bracket
- · You enter 15 percent of hands or more but raise fewer than half of the ones you enter
Stack depth changes the whole question, so a 12 big blind decision is judged against other 12 big blind decisions rather than against your average.
The cash detectors work the same way, with their own thresholds and the same two rules about money and margins, on the leak finder page.
push or fold
Short stack decisions are the part of a tournament that can actually be drilled, because the choice is narrow and the reference is written down.
- · Twenty five shove ranges: five positions at 8, 10, 12, 15 and 20 big blinds, wider the later you sit and the shorter you are.
- · Twenty five defending ranges for the big blind facing a shove, by the position of the player shoving and by stack depth. That is the decision most short stack money runs through, and it was missing entirely until it was added.
- · Twenty of those defending ranges also exist in an ICM version, which is considerably tighter: near a pay jump a call that breaks even in chips is a losing call in money.
- · ICM covers the calling side only, and the pages say so. Half an ICM mode that pretends to be complete is worse than one that names what it covers.
- · The ranges are conventions used in training, not solver output. The source column in the database says exactly that, and so does this sentence.
You can work out what a pay jump does to a specific spot with the ICM calculator, and the glossary entry on ICM explains why chips and money come apart in the first place.
what it will not tell you
Six limits. Most of them exist because the honest answer for tournaments is quieter than the one people expect.
- · Return on buy in needs volume. Under thirty tournaments at a buy in level it says very little, and the app says so next to the number rather than hiding the number.
- · A result file that never arrived means no financial result. Hand histories alone give you frequencies and leaks, but the prize, the field and the finishing place come from the summary file.
- · Tournament leaks never carry a price in euros. Every one of them is reported in chips or as a frequency, on purpose.
- · There is no risk of ruin percentage for tournaments. The usual formula assumes results are roughly normally distributed, and a tournament result is not: it is skewed, with a long tail upward. A number that looks precise and is wrong is worse than no number.
- · Bankroll guidance is therefore stated as buy ins: 200 for large fields, 100 as the common standard, 50 for small fields and sit and go formats, 30 to take one shot at the level above.
- · The push or fold detector measures accuracy against a reference, not money won. It needs at least 50 graded decisions and fires below 80 percent, naming whether you were mostly too loose or mostly too tight.
next
One plan covers cash and tournaments.
You do not pay twice for playing both. The pricing page shows which plan opens the deeper tournament tools, the ICM situations, the full push or fold ranges and the breakdown per buy in level.
questions
Do I need a separate subscription for tournaments?
No. One plan covers cash and tournaments, and the two are measured separately so you can see where you stand in each without paying twice. The deeper tournament tools sit on the higher plan.
Why is there no euro amount on a tournament leak?
Because a chip is not a fixed amount of money. Its value depends on the stage of the tournament and the payout structure, and without a full ICM calculation any conversion is a guess. Chips are used as a gate instead: a pattern that gained chips is never reported as a leak.
Does it know when I was on the bubble?
Only if the result file is there. A hand history gives the blind level, which supports early and middle and nothing more. The bubble, the money and the final table depend on players remaining against the payout structure, and inventing them from a blind level would be a guess.
How many tournaments before my return on buy in means anything?
Thirty is where the app starts trusting it, and even then it is one of several components rather than the verdict. Under that threshold the score is pulled back toward the middle and the number is shown with that caveat attached.
Are the push or fold ranges solver output?
No, and the database says so in its own source column. They are the conventional ranges used in training material, which is a useful reference and not a solved strategy. The ICM variants cover calling a shove from the big blind, not shoving itself.