One method for every record.

eToro accountHyperliquid vaultBacktest

One methodSame costs, same metrics, same score

A public pageLabelled live or backtest

• Live recordsRead directly from the trading platform, such as eToro or Hyperliquid, through read-only access, with the real fees. Owners cannot edit them.
• BacktestsA backtest is hypothetical. OXrun runs the strategy on past prices with a cost model. Owners cannot upload results.
• One methodEvery strategy is calculated the same way: the same cost model, the same metrics and the same robustness score.
• Backtests score lowerThey are weaker evidence than live trading, so a backtest alone cannot reach the top of the robustness score.
• Reference strategiesBuy-and-hold baselines, such as the S&P 500 (SPY), the Nasdaq-100 (QQQ) and a 60/40 stock and bond mix, give every result something to compare with.

The score shows how likely a record is to repeat, not how high it is. It checks five things.

• Statistical evidence

  Is the edge real, given the length of the record? Measured with the probabilistic Sharpe ratio.

• Overfitting risk

  Do results hold on data the strategy was not built on (out-of-sample)? This catches overfitting.

• Consistency over time

  Does it work in rising, flat and falling markets?

• Live evidence

  Real trading counts more. A backtest alone cannot reach the top.

• Implementability

  Do results hold after higher trading costs and a one-day delay?

One score from 0 to 100

The five parts are weighted as below. Without a live record, the maximum is 85.

• Statistical evidence
• Overfitting risk
• Consistency over time
• Live evidence
• Implementability

On every strategy page