How much money do I need to start betting tennis?

Whatever amount you can comfortably lose — that's your working bankroll. With 1u = 1%, a €1,000 bankroll means €10 stakes. Don't start higher than you can absorb 100u of losses on. Most beginners overstake by 3-5×.

Should I increase stake size after a winning streak?

Only if your bankroll has grown by a meaningful amount (e.g. +20%) across a 100+ bet sample. Increasing during a short winning streak is variance chasing — when the inevitable drawdown comes, you'll be at peak stake size and lose more than you should.

Is fractional Kelly worth it vs flat stakes?

Marginally yes IF your probability estimates are well-calibrated. For most bettors with imperfect probabilities, flat stakes are safer with only ~10% lower expected growth. TIPERO uses 0.15-Kelly with hard caps — a balanced compromise.

What's the safest bankroll percentage for a beginner?

1u = 0.5-1% of working bankroll. The lower end (0.5%) lets you absorb 200u of drawdowns. The higher end (1%) is standard for tracked syndicate-style strategies. Above 2% is aggressive and statistically high-risk.

Variance and Bankroll Management for Tennis Bettors

Edge gets you nowhere if variance wipes you out first. Here is the practical math of bankroll survival — drawdowns, units, fractional staking and the rules that keep you in the game long enough for your edge to play out.

Variance is louder than edge in the short run

Even a +5% ROI bettor — extremely sharp by tennis standards — will see 30+ bet losing streaks on a regular basis. Plug a +5% edge into a Monte Carlo simulator with 1u flat stakes:

10,000 simulated 1,000-bet seasons at +5% ROI:
  Median ending bankroll:    +50u
  10th percentile:           -22u   ← you'd think you have NO edge
  90th percentile:          +123u
  Worst run-up drawdown:    -45u
  Probability of -30u DD:   ~38%

Read those numbers slowly. With a real, sustained edge, you still have a 38% chance of seeing a 30-unit drawdown inside the season. If your bankroll can't absorb that — meaning, if 30u of drawdown forces you to reduce stake size or stop — your edge can be perfectly real and you can still go broke.

Units, percentages and what bankroll math actually means

"1 unit" is the only sensible way to talk about stake size across people. The convention used by most syndicates and tracked tipsters:

1u = 1% of your working bankroll
   (working = the money you're explicitly willing to lose,
    NOT total savings, mortgage, etc.)

  Conservative:   1u = 0.5% of working bankroll
  Standard:       1u = 1.0%
  Aggressive:     1u = 2.0%   (most blow-up scenarios live here)

If your working bankroll is €5,000 at 1u = 1%, then 1u = €50 and a 1.5u TIPERO pick = €75 placed. A 30u drawdown would dent the bankroll by €1,500 — survivable. The same drawdown at 1u = 2% = €100 stakes wipes €3,000 — you're now stake-restricted whether you want to be or not.

The drawdown survival rule

The single most important rule in bankroll management:

Your bankroll must be able to absorb 3× your worst historical drawdown without forcing you to reduce stake size.

If TIPERO's worst historical drawdown was -28u, you should size your unit so a -84u drawdown still leaves you betting at the same unit. Reason: drawdowns can and will exceed historical worst — you're sampling from a distribution, not viewing the cap.

Flat stakes vs fractional Kelly

Two valid staking philosophies:

Flat stakes — every bet is exactly 1u (or scaled by tier: 1u CORE, 0.5u VALUE, 0.3u LONG/ULTRA). Boring but bulletproof. Variance-minimal.
Fractional Kelly — stake size scales with edge. A 12% EV pick gets a bigger stake than a 3% EV pick. Higher long-run growth IF probability estimates are well-calibrated.

For most bettors with imperfect probability estimates, flat stakes are mathematically safer. Kelly's growth-rate advantage assumes you know your probability exactly — which you don't. TIPERO ships pre-sized Kelly stakes (capped at 1.5u max) using calibrated probabilities, but if you prefer flat 1u everything, you'll still capture most of the edge.

The four bankroll-killer mistakes

Sizing relative to the last bet — "I lost €200 yesterday, I'll bet €400 today to make it back." Classic Martingale path to zero. Stakes are sized off your current bankroll, never off recent results.
Increasing unit size during winning streaks — feels good, exposes you to the inevitable mean-reversion drawdown at peak stake. Re-size only after the bankroll has demonstrably grown over a 100+ bet sample.
Dropping stake size during drawdowns — if your bankroll genuinely shrinks (e.g. -20%), then yes scale stakes down proportionally. But "I'm down 8 bets, I'll bet less" within a normal variance window is just chasing the variance.
Treating betting bankroll as life money — if you can't lose all of it without affecting rent, food or mental health, the size is wrong. Period.

How TIPERO models variance internally

The pipeline runs Monte Carlo simulations over the prediction log every week — sampling 1,000 random subsets of historical bets to bracket the realistic range of season outcomes. The weekly Truth Report (sent to admins) flags any tier whose realised variance exceeds modelled variance — that's our drift detector for the bet selection rules.

Bottom line

Edge is necessary but not sufficient. Bankroll survival is what lets your edge actually compound. Set a unit size you could lose 100u of without flinching, stake flat or fractional Kelly, never chase, and trust the law of large numbers. That's the boring truth that separates profitable bettors from the broken ones.

See TIPERO's pre-sized Kelly stakes →

Variance and Bankroll Management: The Math Behind Survival