September 23, 2026
Esports betting content almost always starts with a pick. This article starts one step earlier, with the price — because the price is the only part of a bet that is fully knowable before the match, and it decides how much a correct opinion is actually worth.
Two numbers do most of the work: the probability implied by the odds, and the margin the bookmaker has built into them. Both take about ten seconds to calculate, and once you can, most of the tipping industry becomes readable from the outside.
From decimal odds to probability
A decimal price is a payout multiplier. Divide one by it and you get the probability the market is charging you for: 2.00 implies 50%, 1.50 implies 66.7%, 3.40 implies 29.4%. Nothing subtle happens there. The subtlety is that all the implied probabilities in a market add up to more than 100%, and the excess is the bookmaker’s margin — the overround.
Margin = (sum of implied probabilities) − 1. Three worked examples from the same imaginary series:
| Market | Prices | Sum of implied probabilities | Margin |
|---|---|---|---|
| Match winner, balanced | 1.90 / 1.90 | 52.6% + 52.6% = 105.3% | 5.3% |
| Match winner, shaded | 1.85 / 1.95 | 54.1% + 51.3% = 105.3% | 5.3% |
| Correct score, three-way | 2.40 / 3.60 / 2.55 | 41.7% + 27.8% + 39.2% = 108.7% | 8.7% |
The second row is the one worth staring at. The overround is identical to the first, but it has been moved onto one side. If the match really is a coin flip, backing at 1.95 costs 2.5% of the stake and backing at 1.85 costs 7.5% — same market, same margin, three times the price depending on which team the tip names.
The third row shows what happens the moment a market is derived from another. The plain match winner carries 5.3%; slicing the same series into correct scores carries 8.7%. Map handicaps, map totals and first-blood props all sit in that second category, and the tips that sound cleverest are usually written on them. The break-even hit rate follows directly from the price: at 1.90 you need 52.6% of your bets to land merely to stand still, at 1.85 you need 54.1%. If any of these terms are new, the glossary defines them in a line each.
So the first question to ask of any selection is not whether the reasoning is good. It is what the whole market costs, and whether the tip is on the cheap side of it or the expensive one.
Closing line value: the cheapest evidence there is
The closing price is the last price a market shows before it settles, and it is the best public estimate of the true probability — every opinion, every late roster announcement and every large bet has been folded into it. Beating it consistently is the only cheap sign that a method contains information. Losing to it consistently is proof that it does not, no matter what the profit column says.
Remove the margin before comparing. Suppose the market closed at 1.85 / 1.95. Implied: 54.1% and 51.3%, summing to 105.3%. Divide each by 1.053 and you get 51.3% and 48.7%; the fair price on the side you backed is 1 ÷ 0.487 = 2.05. If you took 2.10, you beat the fair close by 2.4%.
That is a whole tracking system, and it fits in six columns:
| Date | Market | Price taken | Fair close | CLV | Result |
|---|---|---|---|---|---|
| 14 Mar | Team A to win the series | 2.10 | 2.05 | +2.4% | Lost |
| 16 Mar | Under 2.5 maps | 1.80 | 1.90 | −5.3% | Won |
| 21 Mar | Team C +1.5 maps | 1.62 | 1.55 | +4.5% | Won |
| 22 Mar | Team D map 1 | 2.30 | 2.45 | −6.1% | Lost |
Note what the log shows after four bets: one of the winners had terrible closing-line value and one of the losers had good value. Results take thousands of bets to say anything; CLV starts saying something after a few dozen, because it does not depend on whether the ball bounced your way. A tipster who will not publish the price taken alongside the selection has removed the only column that could have been checked.
Reading a tipster or a betting blog
Start with sample size, because it quietly settles most arguments. Take a bettor with no edge whatsoever, flat-staking one unit at around 2.00. After 50 bets, ordinary luck alone leaves them somewhere within about seven units of break-even roughly two thirds of the time, and within fourteen units nineteen times out of twenty. A record of plus twelve units over 50 bets is therefore entirely consistent with having no skill at all. Even after 500 bets, the noise in the measured yield is still around four percentage points — which is larger than almost any real edge. This is not a reason to distrust everyone; it is a reason to treat any short record, good or bad, as uninformative.
Then check the framing of the numbers. Yield is profit divided by total staked; return on bankroll is profit divided by the money at risk. A 5% yield on 1,000 units of turnover is 50 units of profit — on a 100-unit bankroll that is also a 50% return on capital, and you can guess which of the two figures ends up in the headline. Both are honest on their own; quoting one while implying the other is not.
| Claim | What evidence would settle it |
|---|---|
| 78% strike rate last month | The prices. A 78% strike rate at 1.25 loses money; 40% at 3.00 makes it. A hit rate without odds is not a result. |
| Plus 180 units this season | Unit size, total staked and the full bet list. 180 units on 6,000 staked is a 3% yield; on 400 staked it is a claim that needs a lot of proof. |
| Verified record | Who verified it, whether every selection was timestamped before the market closed, and whether the losing months are still on the page. |
| Proven system | Out-of-sample results and closing-line value. A backtest is a description of the past, not a test. |
| Free tips, no catch | How the site earns: affiliate commission, subscriptions or advertising. All three are legitimate; the undisclosed version is the problem. |
Selective reporting is the failure mode to watch for, and it rarely looks like lying. A channel posts ten selections, deletes the three that lost, and the remaining screenshots are all genuine. Ask whether the record is a list you can count, or a gallery you are shown. Business models matter for the same reason: a subscription earns whether the tips win or not, and affiliate revenue — which is how this site is funded — is earned when a reader signs up, not when a bet lands. Worth saying plainly while we are on the subject: the brand name this site carries, Betiyo, has no working platform behind it, since betiyo.com is a parked domain offered for sale. There are no bonus terms or payouts here to verify, and nothing on this page is a tip.
What makes esports odds different
Everything above applies to any sport. Four things make esports markets behave with less mercy than a football coupon:
- Patches age your sample. A balance patch every few weeks rewrites what works, so a team’s last thirty maps may describe a game that no longer exists. Historical data goes stale in weeks, not seasons — which is also why any model needs the current-patch subset, however small, as described in the esports betting strategy guide.
- Rosters move constantly. Stand-ins and transfers are announced hours before matches, and a famous organisation with two new players is a new team wearing an old logo.
- Map and veto markets cost more. Pick-and-ban derivatives carry the wider margins shown in the first table, and they are exactly where a tip sounds most expert.
- Small tournaments have thin liquidity. Lower limits, wider prices, more suspensions, and a closing line built on far less money — so CLV measured on an obscure qualifier means less than CLV on a major.
In-play markets add the delay problem on top of all of this: your stream runs 30 to 120 seconds behind the server while the book prices off the server, which is covered in detail in the article on live esports betting and stream delay. A tipster posting live selections without addressing that is either trading on a feed you do not have, or has not thought about it.
Before the next bet. Stake in units, one per cent of the bankroll, and keep the size the same whether the last bet won or lost. Set a session limit and a monthly loss limit before opening the site, and treat them as spent money, not a target to recover. Never bet with borrowed money, a credit card or funds meant for something else. Record the price you took — a record without prices cannot answer the only question worth asking. If betting has stopped being entertainment, a national support service will talk it through free and in confidence: Peluuri in Finland, GamCare in the United Kingdom, or the equivalent where you live.
Frequently Asked Questions
How do I calculate a bookmaker margin?
Convert every price in the market to an implied probability by dividing one by the decimal odds, add them up, and subtract 1. Two-way odds of 1.90 and 1.90 give 52.6% + 52.6% = 105.3%, so the margin is 5.3%. Three-way and derived markets such as correct score usually run wider — 8% to 10% is common — which is why the same opinion expressed on a map handicap costs more than on the match winner.
What is closing line value and why does it matter?
Closing line value is the difference between the price you took and the fair price at which the market closed, once the margin is removed. It matters because the closing price is the best public estimate of the true probability, so beating it repeatedly is evidence that your method contains information, while losing to it repeatedly is evidence that it does not. Results need thousands of bets to be meaningful; CLV starts to be informative after a few dozen.
How many bets does a tipster record need before it means anything?
Far more than is usually shown. Flat-staking around even money, a bettor with no edge at all finishes within roughly seven units of break-even after 50 bets about two thirds of the time, so plus twelve units over 50 bets proves nothing. Even 500 bets leaves an uncertainty of about four percentage points in the measured yield. Judge a record by whether prices are published before events start, whether losing runs are still visible, and whether the staking is disclosed in units.
Odds are a price list, not a prediction. Read the margin first, record the price you took second, and let the closing line tell you over time whether anything else you are doing is working.
Model-based tips deserve the same test: what AI can and cannot predict in betting.