Downswings, chance of a loss, ROI spread and bankroll — thousands of simulated samples
Paid places 75, min cash $14.76, 1st place $860. At 20% ROI you cash in 20% of tournaments (no edge: 15%) and win 1 in 375.
| Percentile | Profit | BI | ROI |
|---|---|---|---|
| 5% | … | … | … |
| 25% | … | … | … |
| 50% | … | … | … |
| 75% | … | … | … |
| 95% | … | … | … |
A downswing is the deepest drop from a bankroll peak to the following low within the sample. A stretch without a new high counts the tournaments you went without beating your previous best.
| Place | Prize | Buy-ins | % of pool | Your chance |
|---|---|---|---|---|
| 1 | $860 | 86 | 19.12% | 0.27% |
| 2 | $593 | 59.3 | 13.19% | 0.27% |
| 3 | $426 | 42.6 | 9.47% | 0.27% |
| 4 | $316 | 31.6 | 7.03% | 0.27% |
| 5 | $241 | 24.1 | 5.37% | 0.27% |
| 6 | $189 | 18.9 | 4.2% | 0.27% |
| 7 | $151 | 15.1 | 3.36% | 0.27% |
| 8 | $123 | 12.3 | 2.74% | 0.27% |
| 9 | $102 | 10.2 | 2.28% | 0.27% |
| 10–18 | $54.21 | 5.4 | 1.2% | each 0.27% |
| 19–27 | $25.79 | 2.6 | 0.57% | each 0.27% |
| 28–45 | $17.83 | 1.8 | 0.4% | each 0.27% |
| 46–63 | $15.41 | 1.5 | 0.34% | each 0.27% |
| 64–75 | $14.86 | 1.5 | 0.33% | each 0.27% |
For place ranges the average prize and the chance of each single place are shown. The structure is generated by the model and close to typical GGPoker and PokerStars payouts; a specific tournament may differ.
Poker variance is the gap between what you should win on average and what your graph actually shows. It matters in cash games and spins, but in multi-table tournaments (MTTs) it is extreme: only ~15% of the field cashes, and most of the prize pool goes to the final table. A player with a good ROI loses the buy-in in most tournaments, min-cashes now and then, and makes one deep run every few hundred tournaments that pays for everything else.
That is why a hundred or two tournaments say almost nothing about skill, and a 100–200 buy-in downswing is normal even for a strong regular. This MTT variance calculator simulates thousands of possible samples and shows what results are normal at your ROI.
ROI is your average profit per buy-in: (winnings − buy-ins) ÷ buy-ins. At a 20% ROI and a $10 buy-in you make $2 per tournament on average. ROI is net of rake: with 10% rake a player with no edge at all has an ROI of −10%. 10–30% is a good ROI in mid-sized online fields; micro stakes allow more, high stakes noticeably less. Use the ROI from a big sample (1,000–2,000+ tournaments), and if in doubt, shade it down.
ITM(in the money) is the share of tournaments in which you cash. The organiser sets the share of paid places: usually 15% of the field, sometimes 10–12% (steeper structures) or 20%. In the calculator you pick the tournament’s paid percentage, and the model works out your personal ITM: at a 20% ROI it rises from about 15% to 20%.
Expected profit grows in proportion to the number of tournaments N, while the spread grows only with √N. Over a short sample luck dominates; over a long one your true ROI shows through. Example for 500-player fields at a 20% ROI:
| Tournaments | Chance of a loss | Observed ROI, 90% of the time | Typical downswing |
|---|---|---|---|
| 100 | 44% | −60% to +144% | 37 BI |
| 500 | 26% | −23% to +68% | 93 BI |
| 1,000 | 16% | −11% to +55% | 126 BI |
| 5,000 | ~1% | +6% to +35% | 225 BI |
| 10,000 | <0.1% | +10% to +31% | 273 BI |
Note that the longer the sample, the deeper the worst downswing inside it — you simply have more time to hit one. Even over 10,000 tournaments the typical drop exceeds 250 buy-ins.
$10 buy-in, 10% rake, 15% paid. The expectation is +200 buy-ins ($2,000). Here is what the simulation shows:
The median player goes ~430 tournaments without a new high, and the longest run without a cash is about 25 tournaments (37+ for 5% of players). At a 10% ROI it gets much worse: a 33% chance of being down after 1,000 tournaments and a typical downswing of 153 buy-ins.
The bankroll comes from the risk-of-ruin formula B = −ln(risk) × σ² / (2μ), where μ is the mean profit and σ the standard deviation per tournament. At a 20% ROI and 15% paid:
| Field | 5% risk | 1% risk | Typical downswing over 1,000 MTTs |
|---|---|---|---|
| 100 | 95 BI | 146 BI | 65 BI |
| 200 | 157 BI | 241 BI | 88 BI |
| 500 | 302 BI | 464 BI | 126 BI |
| 1,000 | 492 BI | 757 BI | 161 BI |
| 5,000 | 1,519 BI | 2,335 BI | 251 BI |
Variance is the spread of results around your expectation. In MTTs almost all of the profit comes from rare deep runs, so even a strong player can be down for hundreds of tournaments in a row. A poker variance calculator shows how wide that spread is for your ROI, field size and sample.
Enter the buy-in and rake, the average field size, your ROI and the number of tournaments. The calculator builds a payout structure, fits finishing-place probabilities to your ROI and simulates thousands of samples. You get the chance of being down, typical and bad downswings, the range of outcomes and the bankroll for your chosen risk of ruin.
ROI (return on investment) is your average profit per buy-in: (winnings − buy-ins) ÷ buy-ins. A 20% ROI means you make $2 on average for every $10 buy-in. ROI is net of rake: a player with no edge has an ROI equal to minus the rake.
ITM (in the money) is the share of tournaments in which you cash. Most MTTs pay 10–20% of the field, usually around 15%. A winning player cashes more often than average: about 20% of the time instead of 15% at a 20% ROI under the standard model.
A lot. In 500-player fields, after 1,000 tournaments a player whose true ROI is 20% will see an observed ROI anywhere between −11% and +55% 90% of the time. After 5,000 tournaments it narrows to +6% to +35%. The bigger the field, the longer the sample you need.
For 200–1,000 player fields at 10–20% ROI, 150–500 buy-ins is the usual range; for large fields, 1,000 or more. The exact figure depends on your ROI, field size and the risk of ruin you accept — the calculator uses B = −ln(risk) × σ² / (2μ).
The model assumes your ROI is known and constant. In reality ROI is often lower than you think, shifts with form and stakes, and real structures with re-entries and bounties behave differently. If you are unsure about your ROI, also run it 5–10 points lower.
Like most variance calculators (PrimeDope, for example): the chance of every paid place is raised by the same factor until the average payout matches your ROI. The top-heavy mode moves the edge toward deep runs, which increases variance. Late registration and re-entries are ignored.