Doinp

PROPHET

Quant model, tracked in the open.

Each model targets a specific market and ships with its full historical record. No cherry-picking, no after-the-fact edits — the cards below update as cycles settle.

PROPHET

Pre-match
PROPHET
Flagship · multi-sport
Pre-match

How it works

PROPHET watches the sharpest reference market across every sport it covers. When the price on a side drops, that's signal: the smart money has moved. PROPHET flags those drops in real time, showing where the move appears at slower books before they re-price.

What you get

  • Real-time alerts on dropping prices across all sports covered.
  • Suggested softer books that haven't adjusted yet.
  • A public, append-only record of every pick the model ever flagged.
Equity curve Cumulative profit
7D
+104.6u
+23.8%
30D
+210.9u
+24.9%
1Y
+206.4u
+24.1%
886
Picks
38.92%
Win rate
3.40
Avg odds
+206.4u
Profit
+24.08%
ROI
+12.64%
CLV
PROPHET SOCCER
Soccer-only · same engine
Pre-match

How it works

Same dropping-odds engine as PROPHET, using OddsNotifier, narrowed to soccer leagues. PROPHET SOCCER only runs on weekends — Friday, Saturday and Sunday — when the soccer calendar is busiest. The model tunes its thresholds and timing windows specifically for that pace.

Coverage

  • Almost every soccer league the reference market covers.
  • Only Match winner.
  • Track record kept separately from the main PROPHET stream.
Equity curve Cumulative profit · by weekend (Fri–Sun)
7D
+3u
+4.2%
30D
+11u
+5.6%
1Y
+37u
+5.1%
612
Picks
38.92%
Win rate
2.18
Avg odds
+37u
Profit
+5.1%
ROI
+1.9%
CLV

Monte Carlo simulation

Updated monthly June 2026

Advanced betting simulation

A drawdown Monte Carlo simulation that stress-tests PROPHET's edge — showing the efficacy of the system from a few basic metrics: how likely the results are skill rather than luck, and the drawdowns to expect along the way.

System & significance

Profit+142u

Total units won over the sample. One unit = one standard stake, so this is the bankroll growth at flat staking.

Hitrate52.9%

Share of settled picks that won. At odds above even-money the model profits without needing a majority of winners.

Profit per bet (μ)0.084u

Average units won per pick — the model's raw edge on a single bet, before variance.

Yield std. dev. (σ)0.89

How widely single-bet results scatter around that average — the bet-to-bet volatility you have to ride out.

T-Statistic3.88

How many standard errors the profit sits above zero. The higher it is, the harder the edge is to write off as noise.

P-Value0.01%

Chance a no-edge system would match these results by luck alone. Lower is stronger evidence of real skill.

1-in-X9,600

The p-value in plain terms — luck this good would show up roughly once in this many identical samples.

Drawdown & risk

Expected max drawdown (EMDD)18.2u

The worst peak-to-trough dip to plan for over a sample this size. Size your bankroll to survive it comfortably.

Montecarlo avg. max drawdown (XMDD)19.0u

Average worst dip across thousands of simulated re-runs of the system — a robustness check on the figure above.

Median max drawdown15.4u

The midpoint worst dip — half of simulated runs stay shallower than this, half go deeper.

Prob. of max drawdown ≥ 50u1.8%

Odds of suffering a brutal 50-unit drop at some point in the run. Low here means tail risk is contained.

Prob. of yield ≥ 099.9%

Odds the system finishes the sample in the green rather than break-even or down.

Profit / EMDD7.8

Reward-to-risk against the expected drawdown. Higher means more profit earned per unit of pain.

Profit / XMDD7.5

The same reward-to-risk ratio, measured against the simulated average drawdown.

System & significance

Profit+37u

Total units won over the sample. One unit = one standard stake, so this is the bankroll growth at flat staking.

Hitrate48.6%

Share of settled picks that won. At odds above even-money the model profits without needing a majority of winners.

Profit per bet (μ)0.051u

Average units won per pick — the model's raw edge on a single bet, before variance.

Yield std. dev. (σ)0.94

How widely single-bet results scatter around that average — the bet-to-bet volatility you have to ride out.

T-Statistic1.79

How many standard errors the profit sits above zero. The higher it is, the harder the edge is to write off as noise.

P-Value3.7%

Chance a no-edge system would match these results by luck alone. Lower is stronger evidence of real skill.

1-in-X27

The p-value in plain terms — luck this good would show up roughly once in this many identical samples.

Drawdown & risk

Expected max drawdown (EMDD)11.4u

The worst peak-to-trough dip to plan for over a sample this size. Size your bankroll to survive it comfortably.

Montecarlo avg. max drawdown (XMDD)12.1u

Average worst dip across thousands of simulated re-runs of the system — a robustness check on the figure above.

Median max drawdown9.3u

The midpoint worst dip — half of simulated runs stay shallower than this, half go deeper.

Prob. of max drawdown ≥ 50u6.2%

Odds of suffering a brutal 50-unit drop at some point in the run. Low here means tail risk is contained.

Prob. of yield ≥ 096.3%

Odds the system finishes the sample in the green rather than break-even or down.

Profit / EMDD3.2

Reward-to-risk against the expected drawdown. Higher means more profit earned per unit of pain.

Profit / XMDD3.1

The same reward-to-risk ratio, measured against the simulated average drawdown.

Figures recompute on the 1st of each month from the full settled track record — they are not edited by hand between runs. As a rule of thumb, a p-value below 0.1% and a Profit/EMDD above 5 indicate a strong, low-variance edge.