+03 Moscow, Volgograd
Profile
T
Eskhata 2:1 Istaravshan · 66'Enisey 0:1 Arsenal Tula · 28'Guadalajara Chivas 3:0 Club Tijuana · 45'Port FC 0:1 Vissel Kobe · 45'Karvan 2:0 Zaqatala · 28'
Menu
Statistics sections16.09.2026
Matches
Schedule, results, lineups
550matches today
TodayTomorrow
Teams
Form, streaks, home and away
24810teams
OverviewSeason
Leagues and Cups
Tables, tournament trends
1240tournaments
Top-5All
Referee
Cards, penalties, strictness
19380referees
RatingBy league
Analytics tools16.09.2026
Dropping odds
Line movement and volumes
397639signals today
TodayTomorrow
Day Analysis
Summary of the day's matches
550matches in the overview
OverviewSeason
AI-match predictions
ML model probabilities
predictions with value
ValueAll
Tools
Referees, calendar, load, scenarios
4tools
Referee modelBuilder
Matchlists
Export data to Excel
XLSXexport format
DownloadHistory
Lesson 3 of 115 min

Bankroll and variance

Why stake size matters more than match selection, what a losing run looks like inside a winning strategy, and how many bets it takes before the result means anything.

This is the dullest lesson in the course and the most useful one. Match selection decides whether you win a particular bet. Stake size decides whether you stay in the game long enough for selection to matter at all.

Bankroll and unit size

Your bankroll is the money set aside for betting, as a whole and separately from everything else. Your unit is the share of that bankroll placed on a single bet, usually 1–3%. You do not stake “a thousand”, you stake “two percent”: as the bankroll grows the stake grows with it, and as it falls the stake shrinks automatically.

Why a percentage rather than a fixed sum. With a bankroll of 50,000 and a stake of 5,000 (10%), seven consecutive losses are enough to wipe out more than half of it. Seven losses in a row is not a disaster and not bad luck — it is an ordinary event that will happen to you several times a year.

How long losing runs get

Suppose you bet at odds of 2.00 and win 55% of your bets — a very strong result, clearly ahead of the market. Even then:

  • a run of 5 losses appears roughly once every 50 bets;
  • a run of 8 appears roughly once every 400 bets;
  • losing months will occur regularly.

This is variance: the spread of results around their average. It does not mean the strategy is bad. It means that over a short stretch, telling a good strategy from a bad one is impossible in principle.

How many bets you need

Fewer than a hundred settled bets tell you almost nothing. Over 30 bets a random person with a coin regularly shows +20% ROI — and just as regularly −20%. This is exactly why the tipster rating has a minimum bet count: without it the top places would belong to people who made five lucky picks and stopped.

One metric does work earlier — CLV: it compares the odds at publication with the odds at kick-off and measures pricing accuracy rather than luck. It reveals an author’s quality over dozens of bets rather than hundreds.

What to remember about money

  • Betting is not a source of income but a negative-expectation game by default: the margin works against you from the very first bet.
  • A bankroll is an amount you could lose and be fine. If that is not true, the amount is wrong.
  • Chasing losses by raising your stake is the fastest known way to reach zero.

More on ROI and how to count profit correctly is in a separate article.

Next

The basics are done. From here on it is about the site itself: where its data comes from and how it differs from what others show.

What to remember

Even a correct strategy goes through long losing runs. A fixed percentage of the bankroll protects you from them; a hundred bets is the minimum sample worth judging at all.