How is PBO computed from a returns matrix?

Combinatorially Symmetric Cross-Validation (CSCV): split the returns matrix into S equal subsets and form in-sample / out-of-sample partitions. The number of partitions grows fast, so the tool samples up to 500 of them, measures how often the in-sample top performer falls below the out-of-sample median, and divides by the number sampled. It reports the resulting PBO, not the per-partition ranks.

What's the Deflated Sharpe Ratio?

Lopez de Prado's adjustment to standard Sharpe that accounts for skew, kurtosis, and the number of trials run. A DSR above 0.95 (95% confidence) is the rough threshold for 'this Sharpe is unlikely to be from random search'. The tool reports both raw and deflated.

What's a common mistake when using Backtest Overfitting Score?

Hiding the trial count. PBO + DSR assume the user reports every parameter combination tried (including the ones killed in the IDE). Under-reporting trials underestimates overfitting.

What is the most common way PBO gets misread?

Reading PBO as a p-value. PBO is a probability that out-of-sample rank is below the median across trials, not a hypothesis test. Use it as a relative risk signal.

Backtesting & Validation Calculator Guide

How to use Backtest Overfitting Score

From an uploaded backtest trade log, it computes the Probability of Backtest Overfitting (PBO), Deflated Sharpe Ratio (DSR), and the probability of skill (PSR) so you can quantify how much of an apparent edge is real versus selection bias.

5 STEPSPublished May 12, 2026Live Content

By AI Fin Hub Research · AI Fin Hub Team

Best Next MoveCalculators

Backtest Overfitting Score

Upload a backtest trade log and compute Probability of Backtest Overfitting (PBO), Deflated Sharpe Ratio, and the odds your edge survives live trading.

CalculatorOpen ->

On This Page

Overview 5 steps Scenarios FAQ

What It Does

Use the calculator with intent

Quants and retail backtesters who tried more than a handful of parameter combinations and need to know whether the best one is genuinely skillful or just lucky.

Interpreting Results

PBO above ~0.5 means more than half of trials would have outperformed the chosen strategy out-of-sample — the strategy is more likely overfit than skillful. Deflated Sharpe corrects the headline Sharpe for the number of trials; a positive value is the real edge after the selection penalty.

Input Steps

Field by field

1

Upload data

Upload your returns matrix: rows = periodic returns (typically daily), one column per strategy variant. Annualization assumes daily rows (×√252). Minimum 16 variants for a stable PBO estimate.
2

Set parameters

Set the number of CSCV partitions (default 16). More partitions = more stable estimate, longer runtime.
3

Read outputs

Read PBO (probability of backtest overfitting) — values above 0.5 mean the in-sample winner is likely to underperform out-of-sample.
4

Read outputs

Read Deflated Sharpe Ratio alongside. PBO measures relative overfitting; DSR measures absolute statistical significance after multiple-testing penalty.
5

If

If PBO > 0.5 or DSR < 0.95, treat the backtest as curve-fit. DSR here is a probability in [0,1], so 0.95 is the 95%-confidence bar. Reduce variant count, lengthen sample, or test on truly fresh data before live deployment.

Common Scenarios

Use realistic starting points

Single backtest, no parameter sweep

Trade log rows

500

Trials tried

In-sample Sharpe

1.4

The tool requires at least two strategy columns to run CSCV. With a trivially small trial count the expected-max haircut is near zero and DSR (a probability in [0,1]) is close to its undeflated value — near PSR against a zero benchmark.

Heavy parameter sweep

Trade log rows

500

Trials tried

200

In-sample Sharpe

2.1

The deflated Sharpe estimate falls well below the in-sample 2.1 once the trial count is honest, and DSR confidence (a probability in [0,1]) drops below 0.95; PBO above 0.5 means the chosen parameter set probably came from luck.

Try These Tools

Run the numbers next

PlaygroundsCalculator

Walk-Forward Validator

Upload a returns CSV. Rolling or expanding IS/OOS windows, per-window Sharpe, walk-forward efficiency, and a concatenated OOS equity curve. Catches regime.

Launch toolOpen ->

CalculatorsCalculator

Deflated Sharpe Ratio Calculator

Bailey & López de Prado deflated Sharpe — corrects observed Sharpe for selection bias across K trials. Reports deflated Sharpe, PSR (probability of skill).

Launch toolOpen ->

GeneratorsCalculator

Synthetic Market Data Generator

Generate synthetic price series — geometric Brownian motion, GARCH(1,1) with volatility clustering, regime-switching bull/bear, or copula-linked.

Launch toolOpen ->

FAQ

Questions people ask next

The short answers readers usually want after the first pass.

PBO is the probability that the strategy with the best in-sample Sharpe ranks below median out-of-sample. Bailey, Borwein, Lopez de Prado, and Zhu (2017) introduced the metric. A PBO above 0.5 means the in-sample winner is more likely to underperform than outperform in production — i.e., the backtest is more curve-fit than predictive.