V8 Predicts

Bankroll and Risk

Staking as a survival problem. Concepts only, and no recommendation about what to risk.

Lesson 1

Surviving the drawdown

After this you can: Size bets so that a normal losing stretch is survivable.

Even a profitable approach has long losing runs. That is variance, not failure, and the arithmetic is unforgiving: risking a quarter of a bankroll per bet makes ruin likely within a handful of losses regardless of how good the selections are.

A defined bankroll and a small unit size turn each bet into a fixed fraction of a known total. This is what makes a bad month uncomfortable rather than terminal, and it is a separate skill from choosing bets well.

Drawdown is the number that matters here, not average return. The average is what you get if you are still betting at the end, and the drawdown determines whether you are.

V8 publishes performance in units and does not tell you what to stake. That is a deliberate boundary, not an omission.

Key takeaways

  • Stake size, not selection quality, determines whether you survive a losing run.
  • Drawdown is the figure that decides whether a process is livable.
  • V8 reports in units and makes no staking recommendation.

Terms used: Bankroll, Unit size, Drawdown, Risk of ruin

Lesson 2

Correlation hides your real exposure

After this you can: Spot when several bets are really one bet.

Bets that look independent often resolve on the same underlying result. Backing a team on the moneyline, on the spread, and on a player total from that team is largely one bet on that team playing well, staked three times.

That matters because a staking plan assumes independence it does not have. Three one-unit bets that all depend on the same outcome are a three-unit position, and a bankroll can take a much larger hit than the plan intended.

The habit worth building is to ask what single result would make every open position lose at once. If the answer covers most of your exposure, you are less diversified than the number of bets suggests.

Key takeaways

  • Correlated bets multiply both upside and downside.
  • Count exposure by underlying outcome, not by number of bets.
  • Ask which single result would sink every open position.

Terms used: Correlation, Concentration, Bankroll, Unit size