How Long Does a Surebet Last? Data from 10,000 Surebets
The median surebet lasts 2 seconds; 5% survive past a minute. Breakdown by market, bookmaker and ROI from Dozor scanner data for September 6, 2026.
Half of all surebets last two seconds or less. The Dozor scanner recorded when 10,000 surebets on esports and prediction markets appeared and disappeared over 20 hours on September 5–6, 2026. The median lifetime was 2.1 seconds, 39% of surebets disappeared within two seconds, 5% lasted longer than a minute, and 1% longer than five minutes. Below is the full distribution and what follows from it.
Methodology
The Dozor server reads bookmaker lines simultaneously and recalculates every surebet on each cycle. When a surebet can no longer be found, the history records when it first and last appeared, its peak ROI, the market type and the books involved. Sample window: from 21:54 UTC on September 5 to 17:48 UTC on September 6, 2026 — exactly 10,000 records, 27 bookmakers and exchanges, 342 distinct pairs.
Lifetime distribution
| Lifetime | Share of surebets |
|---|---|
| under 2 seconds | 39% |
| 2–10 seconds | 31% |
| 10–30 seconds | 17% |
| 30–60 seconds | 8% |
| 1–5 minutes | 4% |
| over 5 minutes | 1% |
Seven surebets out of ten vanish within ten seconds. In that time you need to spot the surebet, open two bookmakers, calculate the stakes and place two bets. That can't be done by hand, which is exactly why arbers use scanners, and why scanners compete on update latency.
By market type
| Market | Surebets | Median lifetime | Over a minute |
|---|---|---|---|
| Handicap | 3,299 | 2.0 s | 3% |
| Winner | 2,519 | 2.1 s | 5% |
| Total | 1,964 | 2.0 s | 3% |
| Home team total | 484 | 2.0 s | 4% |
| Round handicap | 397 | 2.1 s | 5% |
| Away team total | 344 | 2.0 s | 1% |
| Total kills | 271 | 16.3 s | 18% |
All the main markets are equally short-lived. There's one exception: total kills in Dota 2 and LoL. It lasts eight times longer, and one in five surebets on it lives longer than a minute. It's a market that bookmakers update less often and price worse than the match result and the handicap.
ROI and lifetime
Intuition says fat surebets disappear faster. The data doesn't bear that out.
| Peak ROI | Surebets | Median lifetime | Over a minute |
|---|---|---|---|
| under 2% | 3,374 | 2.1 s | 7% |
| 2–5% | 3,589 | 2.1 s | 5% |
| 5–10% | 2,292 | 2.0 s | 3% |
| over 10% | 745 | 4.0 s | 3% |
Lifetime barely depends on ROI. Surebets above 10% live slightly longer by median but survive a minute less often. Such values usually mean one of the books hasn't updated its line after something happened in the match, and the bet on it is likely to be voided or settled at odds of 1.
Which pairs last longer
Among the 483 surebets that lasted longer than a minute, the most common pairs were CSGOEmpire and Melbet, 1win and BC.Game, Polymarket and Winline, Duel and Polymarket, and Duel and Thunderpick. They have one thing in common: different price sources. The prediction market Polymarket, where prices are moved by participants' trades rather than by an analyst, reacts more slowly than traditional bookmakers: the median lifetime of surebets involving it was 4.1 seconds, versus 2.1 overall. Over 24 hours, Polymarket was involved in 1,129 surebets and Kalshi in 187.
What this means in practice
- Scanner latency is decisive. The difference between updating once a second and once every 0.1 seconds is the difference between “saw it after it was gone” and “got the bet in”. With a median lifetime of 2 seconds, a scanner with a 3-second delay mostly shows dead surebets.
- Alerts matter more than the screen. You can't keep a tab open around the clock, and the 5% of surebets that last longer than a minute turn up at any hour. A Telegram alert gets them to you.
- Total kills and prediction markets are the calmest segments. There's time to check and calculate there — and there's also a higher risk that the gap reflects different settlement rules rather than an error.
- A high ROI is a reason to check, not to celebrate. The 10% threshold separates market discrepancies from line errors.
FAQ
Why are so many surebets under two seconds?
It comes down to price timing at two books in live: one has already updated its line after a round, the other hasn't yet. A moment later the second one catches up and the surebet closes.
Does that mean surebets are useless?
No. The 5% lasting over a minute works out to about 25 surebets an hour in this sample, and they're spread across all markets. The point is to hear about them in the first second.
Do prematch surebets last longer?
Yes, prematch lines change less often. A separate measurement by mode is in the article “Live vs Prematch: Where Surebets Last Longer”, and the mode itself is covered on the Prematch surebets page.
How was lifetime measured?
As the difference between the first and last cycle on which the server found the surebet. Surebets found on a single cycle get a lifetime of 0 seconds and fall into the first row of the table.