Tournament tool
Poker ROI Calculator
Tournament ROI, ITM percentage and what they are worth per tournament and per hour. Then the part everybody skips: a confidence interval on that ROI, and how many entries it would take before the figure means anything.
Your results
Time played
Hours per tournament is your time at the table for one entry. If you play several at once, an hour of four tables is still one hour of your life, so enter your total clock hours instead of hours per tournament.
How much to believe it
2.5 to 3.5 buy-ins is the usual range. Estimate. The bigger the field and the more top-heavy the payouts, the higher it goes.
Per-tournament standard deviations are quoted in buy-ins and are estimates compiled from published player discussion, not measured from your results. Use the range rather than the midpoint when you want to know how wrong an interval could be.
Your ROI
Solid winner. A strong result in large fields if it holds up over thousands of entries.
ROI is measured on the total cost of entering, buy-in plus fee, because the fee is money you had to pay to play. A $10 + $1 tournament costs $11, and 100 of them cost $1,100.
In the money
ITM on its own says nothing about profit. In a top-heavy field the money is at the final table, so a player who trades min-cashes for deep runs will show a lower ITM and a higher ROI than one who folds into the money.
Per tournament and per hour
The same 27.27% ROI at $11.00 an entry, against the clock. ROI does not contain the length of a tournament, which is why a big ROI in a long format can be a worse living than a small one in a short format:
| Hours an entry | Entries an hour | Per entry | Per hour |
|---|---|---|---|
| 0.75 | 1.3333 | $3.00 | $4.00 |
| 1 | 1 | $3.00 | $3.00 |
| 2 | 0.5 | $3.00 | $1.50 |
| 4 | 0.25 | $3.00 | $0.75 |
| 8 | 0.125 | $3.00 | $0.37 |
Is it real?
Over 100 entries the interval runs from -31.5% to 86.1%. It spans zero, so this sample cannot tell a 27.3% winner from a break-even player.
This interval is a normal approximation: the standard deviation of the mean is the per-tournament standard deviation divided by the square root of the number of entries. Tournament results are heavily skewed, because most entries return nothing and a rare deep run returns fifty buy-ins or more, so the approximation fails by being lop-sided rather than by being narrow. Simulating skewed results shows where it lands: at small samples almost every interval that misses the true ROI sits entirely above it, because one deep run drags a short sample upwards and carries the whole interval with it. A tidy-looking interval on a few hundred entries therefore flatters you, and samplesNeeded() is the figure to read next.
This is a normal approximation to a one-sided tail, and on skewed tournament results it comes out too small. Simulating a genuinely break-even player with the same standard deviation posts a stretch this good more often than the normal figure says, because the luck that produces a false winning record arrives as one deep run rather than as a steady drift. Read it as a floor on how often luck alone explains your record, not a ceiling, and note that it is not the chance your ROI is fake: that would need a prior on how many players in your pool beat the fee at all.
How many entries you would need
The interval narrows with the square root of the sample, so halving the margin needs four times the tournaments. There is no shortcut: a smaller sample hides the uncertainty rather than removing it.
What ROI does not cover
ROI is measured on the total cost of entering, buy-in plus fee, because the fee is money you had to pay to play. A $10 + $1 tournament costs $11, and 100 of them cost $1,100. Nothing here counts tax, rakeback, a staking deal or the value of a satellite seat, and a positive ROI says nothing about whether your roll survives the swings behind it.
Guide
How to calculate poker ROI, and how much of it to believe
ROI is the figure tournament players quote at each other, and usually the one they have least evidence for. The arithmetic is one line. The hard part is the rest of it: given the entries behind the figure, which true ROIs could have produced it?
The formula
ROI = (total returned − total invested) ÷ total invested, as a percentage
ROI is measured on the total cost of entering, buy-in plus fee, because the fee is money you had to pay to play. A $10 + $1 tournament costs $11, and 100 of them cost $1,100.
So 100 entries at $10.00 + $1.00 cost $1,100.00, not $1,000.00. Collect $1,400.00 and the profit is $300.00: an ROI of 27.27%, or $3.00 a tournament. Against the buy-ins alone it reads 40.00%, the version people quote when they want a bigger number. The fee was 9.09% of every entry and you had to pay it.
| Entries | 100 |
| Cost of one entry | $11.00 |
| Total invested | $1,100.00 |
| Prize money collected | $1,400.00 |
| Profit | $300.00 |
| Profit per tournament | $3.00 |
| ROI on total cost | 27.27% |
| ROI on the buy-in only | 40.00% |
| Fee share of an entry | 9.09% |
A re-entry counts separately, because it cost a separate fee. Booking a three-bullet day as one entry is how a results sheet flatters itself.
ITM percentage answers a different question
Cashing 18 times in 100 entries is an ITM of 18.00%, one cash every 5.6 entries, averaging $77.78 a cash against $14.00 returned per entry played. None of it says whether the play was profitable. In a top-heavy field the money is at the final table, so folding into every min-cash buys a high ITM and a negative ROI. Read ITM as a description of how you play.
ROI and hourly rate are not the same ranking
ROI is profit per pound risked. Hourly is profit per hour of your life. Neither contains the other, because the cost of an entry and the length of a tournament appear in the second and not the first. Three formats, all of them at the same $110.00 entry cost:
| Format | Entry | ROI | Hours each | Per entry | Per hour |
|---|---|---|---|---|---|
| One-hour turbo | $110.00 | 20.0% | 1 | $22.00 | $22.00 |
| Deep-stack MTT, eight hours an entry | $110.00 | 100.0% | 8 | $110.00 | $13.75 |
| 9-man SNG, 45 minutes | $110.00 | 8.0% | 0.75 | $8.80 | $11.73 |
The best ROI there is 100.0%, on the 8-hour format, worth $13.75 an hour. The best hourly is $22.00, on the 1-hour format, from an ROI of 20.0%. The two orderings disagree here. They can disagree as soon as formats differ in cost or in length: only formats that cost the same and take the same time are guaranteed to rank alike.
What counts as a good ROI
| ROI | Reading | Why |
|---|---|---|
| Below -10% | Losing | Worse than the entry fee alone explains, so the leaks are in the play rather than the structure. |
| -10% to 5% | Around break-even | Where most of a field sits once the fee is counted. Game selection and rakeback move this band more than strategy does. |
| 5% to 15% | A small but real edge | A normal figure for a winning regular in big online fields, and plenty at volume. |
| 15% to 30% | Solid winner | A strong result in large fields if it holds up over thousands of entries. |
| 30% to 60% | Strong, or a soft pool | Common in small, soft or live fields. Over a large online sample it is rare. |
| 60% and above | Very high | Usually a small sample with one deep run in it. Check the interval before believing it. |
These bands are rules of thumb for multi-table tournaments, compiled from published player discussion rather than measured. Single-table SNGs sit far lower: 5% to 10% is a good ROI there, because the fee is a bigger share of a smaller prize pool, and their per-tournament standard deviation is under half a large-field MTT's.
The 27.27% above reads as: Solid winner. A strong result in large fields if it holds up over thousands of entries. At $11.00 an entry over 2 hours that is $3.00 an entry, $1.50 an hour.
Is that 27.27% real?
Not on 100 entries. ROI as a fraction is the average profit per entry in buy-ins, so its standard error is the per-tournament standard deviation over the square root of the entries. Large fields run near 3 buy-ins an entry, so at 100 entries the standard error is 30.00% and the 95% interval is 27.27% plus or minus 58.80%:
Over 100 entries the interval runs from -31.5% to 86.1%. It spans zero, so this sample cannot tell a 27.3% winner from a break-even player.
A break-even player with that standard deviation posts a stretch this good about 18.2% of the time, one in 5.5. Volume is the only cure and it works slowly. Quadruple the sample to 400 and the half-width falls from 58.80% to 29.40%, a factor of exactly 2.00, because the square root of four is two. The interval runs -2.1% to 56.7% and still spans zero.
This interval is a normal approximation: the standard deviation of the mean is the per-tournament standard deviation divided by the square root of the number of entries. Tournament results are heavily skewed, because most entries return nothing and a rare deep run returns fifty buy-ins or more, so the approximation fails by being lop-sided rather than by being narrow. Simulating skewed results shows where it lands: at small samples almost every interval that misses the true ROI sits entirely above it, because one deep run drags a short sample upwards and carries the whole interval with it. A tidy-looking interval on a few hundred entries therefore flatters you, and samplesNeeded() is the figure to read next.
Read the 18.2% as a floor rather than a ceiling: on skewed results the normal tail understates how often a break-even player posts a run this good. It is also not the chance your ROI is fake, which would need a prior on your pool.
How many tournaments it actually takes
Invert the half-width and you get the sample size for a margin you choose. At 3 buy-ins, an interval no wider than plus or minus 10 points needs 3,458 entries. Halve the margin to 5 and it needs 13,830: 70 weeks, 1.33 years, at 200 a week. Smaller fields are kinder: 4,979 entries for a 180-man or other mid-field MTT and 2,597 for a 9-man single-table SNG.
The interval narrows with the square root of the sample, so halving the margin needs four times the tournaments. There is no shortcut: a smaller sample hides the uncertainty rather than removing it.
A few hundred entries supports an interval wide enough to hold both a small loss and a career, which makes the figure weak rather than worthless. Cash players meet the same problem in the win rate calculator, and the variance simulator shows the uncertainty as sample runs instead of as an interval.
Pasting your own results instead
Summary totals work when every entry cost the same. When they did not, paste the results and the standard deviation comes from your own list. A run of 12 mixed entries costing $360.00, with $1,003.00 collected, gives 178.61% ROI, $643.00 of profit, 33.33% ITM and $21.43 an hour across 30 hours, on a measured standard deviation of 7.16 buy-ins: well above the 3 preset, because one $780 score is most of the sample.
The interval reads -226.6% to 583.8%, with a lower bound below −100%, which cannot happen: you cannot lose more than you staked, so the approximation has run out of road. A dozen entries does not have an ROI. It has an anecdote, and the list is more use for the spread it measures.
What this model assumes
Entries are treated as independent draws from one distribution: close enough for a mixed schedule, wrong for one tournament re-entered four times in an evening. The standard deviation is an input rather than a measurement: Per-tournament standard deviations are quoted in buy-ins and are estimates compiled from published player discussion, not measured from your results. Use the range rather than the midpoint when you want to know how wrong an interval could be.
Nothing here counts tax, rakeback, staking or the value of a satellite seat. The staking calculator prices a markup deal, the PKO calculator handles bounty equity, and what is rakeback in poker covers the rewards. Nor does a positive ROI mean your roll survives the swings behind it, which is what the bankroll manager and bankroll management are for. And it measures results, not decisions: the late game moves an MTT ROI most, so the ICM calculator, bubble factor and tournament strategy are where this figure gets improved.
Frequently asked questions
How do you calculate poker ROI?›
ROI is (total prize money returned minus total invested) divided by total invested, as a percentage. Invested means the full cost of entering, buy-in plus fee, times the number of entries. 100 entries at $10.00 + $1.00 is $1,100.00 invested, so collecting $1,400.00 is an ROI of 27.27%, or $3.00 of profit per tournament played.
Does poker ROI include the entry fee?›
ROI is measured on the total cost of entering, buy-in plus fee, because the fee is money you had to pay to play. A $10 + $1 tournament costs $11, and 100 of them cost $1,100. Measuring on the buy-in alone turns that example into 40.00% instead of 27.27%, which is why this calculator shows both figures side by side rather than quietly picking the flattering one.
What is a good ROI in poker tournaments?›
In large multi-table fields, 5% to 15% is a real edge and plenty at volume, 15% to 30% is a solid winner if it holds up over thousands of entries, and above 60% is usually a small sample with one deep run in it. These bands are rules of thumb for multi-table tournaments, compiled from published player discussion rather than measured. Single-table SNGs sit far lower: 5% to 10% is a good ROI there, because the fee is a bigger share of a smaller prize pool, and their per-tournament standard deviation is under half a large-field MTT's.
What is the average ROI in poker MTTs?›
Across a whole field it has to be negative, because the fee is taken out of every entry before a chip is dealt. In the worked example the fee is 9.09% of each entry, so the average player in that pool runs at about minus that figure before any difference in skill is counted. Winners are paid out of losers, so a pool where most players hover near break-even is a pool where the fee is doing the losing.
How many tournaments do I need to prove my ROI?›
More than you have played. At 3 buy-ins of per-tournament standard deviation, a 95% interval no wider than plus or minus 10 percentage points takes 3,458 entries, and plus or minus 5 points takes 13,830. Single-table SNGs are kinder because they swing less: the same 5-point margin needs 2,597 entries there. The sample grows with the square of the precision you want, so halving the margin costs four times the volume.
Is my 40% ROI real if I have only a few hundred tournaments?›
Treat any figure from that sort of sample as an estimate with a very wide interval rather than a fact. Over 100 large-field entries at 3 buy-ins of standard deviation, the 95% interval around a 27.3% result runs -31.5% to 86.1%, and a genuinely break-even player posts a stretch that good about 18.2% of the time. Even 400 entries leaves the interval at -2.1% to 56.7%.
What is a good ITM percentage in poker?›
It depends on the structure, and higher is not automatically better. The 18.00% in the example above, one cash every 5.6 entries, is ordinary in a large field paying the top tenth or so. In a nine-man SNG paying three places it would be poor. ITM describes how often you reach the money and not how much money you reach, so a player who trades min-cashes for final tables shows a lower ITM and a higher ROI.
Is a high ROI better than a high hourly rate?›
They answer different questions and the rankings can disagree. In the guide table above, Deep-stack MTT, eight hours an entry has the best ROI at 100.0% and returns $13.75 an hour, while One-hour turbo returns $22.00 an hour on an ROI of 20.0%. ROI tells you how well you play a format; hourly tells you what playing it is worth. Use ROI to choose what to study and hourly to choose what to play.
How do you turn ROI into a poker hourly rate?›
Multiply your ROI by what one entry costs to get profit per entry, then divide by the hours an entry takes. At 27.27% ROI and $11.00 an entry that is $3.00 per tournament, and over 2 hours an entry it is $1.50 an hour. Hours per tournament is your time at the table for one entry. If you play several at once, an hour of four tables is still one hour of your life, so enter your total clock hours instead of hours per tournament.
Related tools
Is your win rate statistically real? Confidence intervals by sample size.
Tournament downswings from a real payout structure: bands, drawdown, dry runs.
Markup math for buying and selling action — fair prices both ways.
Interesting spots. Solver-grade analysis.
Occasional strategy emails: one interesting hand, broken down with equity, range vs range, and the GTO answer. Free, unsubscribe any time.
Unlimited AI hand reviews and every cheatsheet — PRO from $9/mo