How to Analyze TradingView Backtest Results
TradingView gives you win rate, profit factor, and drawdown. That is a start, not a conclusion. Here is how to read the full story your backtest data is telling you.
Why win rate alone misleads
Most traders open the Strategy Tester, see 65 percent win rate, and call it done. But a strategy that wins 65 percent of the time can still lose money if the average loss is three times the average win. TradingView shows profit factor in the summary panel, but it is easy to scroll past it when the win rate looks good.
I did exactly this with an ES futures strategy. The win rate was 72 percent. The equity curve climbed steadily. I funded a small account and started trading it live. Within two weeks I was down 6 percent. The average losing trade was 4.2 times larger than the average winner. I had been so focused on the high win rate that I missed the broken risk-reward profile sitting right in the data.
A single number cannot describe strategy quality. You need to know how the strategy behaves when it is wrong, how consistently it generates returns, and whether those returns are stable enough to survive real market conditions. Win rate is one data point. It is not the headline.
Here are the six areas you need to examine before you trust any backtest: risk-adjusted returns, distribution shape, simulation stability, time consistency, trade efficiency, and position sizing. The sections below walk through each one.
Risk-adjusted metrics: Sharpe, Sortino, and Calmar ratios
The Sharpe ratio is the most widely used risk-adjusted metric in quantitative finance. It divides your excess return (returns minus the risk-free rate, typically 3 to 5 percent) by the standard deviation of returns. A Sharpe of 1.0 means you earned one unit of return for each unit of total volatility. A Sharpe of 2.0 is exceptional for most strategies. Below 0.5 means you are not being compensated for the risk you are taking.
The Sortino ratio works the same way but only penalizes downside volatility. This is important because upside volatility is fine. Sharp spikes in positive returns should not hurt your score the same way steep drops do. When I first compared Sharpe and Sortino on my own backtest data, the Sortino came back at 1.8 against a Sharpe of 1.2. That gap of 0.6 told me most of the volatility was on the upside, which is a good sign. A small or negative gap means your downside volatility is eating into returns.
The Calmar ratio divides annualized return by maximum drawdown. It answers one direct question: was the return worth the worst drawdown you sat through? A Calmar above 3.0 is strong. Below 1.0 means the strategy is producing weak returns relative to its worst-case loss. Use the Calmar ratio calculator to check your numbers.
A common pattern I see is a strategy with a strong Sharpe but weak Calmar. That happens when the strategy has high volatility but also high returns, making the Sharpe look fine. Yet the maximum drawdown is brutal. These are strategies that work well most of the time and then blow up. The Calmar catches this, the Sharpe does not.
TradingView does not calculate any of these three ratios by default. You have to export your data and run the numbers yourself, or use a tool that computes them automatically from your CSV export.
Distribution and fat tails: your returns do not follow a bell curve
Most statistical models assume returns follow a normal distribution. Trading returns do not. Real strategy results produce fat tails, meaning extreme outcomes happen more often than a normal curve predicts. This single fact breaks many of the standard assumptions traders rely on.
This is where Value at Risk (VaR) and Conditional VaR (CVaR) come in. VaR at 95 percent confidence tells you the worst loss you would expect in 95 out of 100 trades. CVaR, also called Expected Shortfall, goes further and tells you the average loss when you are in that worst 5 percent. Two strategies can have the same VaR but very different CVaR if one has mild tail risk and the other has catastrophic blowoffs.
I checked these on a mean reversion strategy that looked clean in the summary stats. The VaR at 95 percent was acceptable at -1.8 percent per trade. The CVaR was -4.2 percent. That gap meant the bad losses were not just frequent but severe when they hit. The strategy was picking up pennies in front of a steamroller, and CVaR exposed it. You can run your own numbers with the CVaR / Expected Shortfall calculator and the Value at Risk calculator.
Skewness and kurtosis complete the distribution picture. Negative skew means your losing trades are larger than your winning trades. High kurtosis means extreme outcomes, both good and bad, are more common than a normal distribution would predict. A strategy with negative skew and high kurtosis is dangerous, even if the average return looks fine. I rejected a crypto mean reversion strategy with positive average returns because the skew was -1.4 and kurtosis was 7.2. Those numbers meant the strategy would blow up eventually, and it did in out-of-sample testing three months later.
Most traders skip distribution analysis because it feels academic. In practice it catches risks that no single metric can reveal.
A returns distribution histogram with a normal curve overlay makes these patterns visible instantly. When I first plotted my own trade returns this way, I spotted a cluster of outlier losses around -3 standard deviations that a simple average would never flag. Dots on a histogram tell you things tables hide.
Monte Carlo simulation: test whether your edge is real
A single backtest run tells you how your strategy performed in one specific sequence of market conditions. Monte Carlo simulation runs hundreds or thousands of versions with randomized trade sequences. It answers the question: if history repeated itself differently, would the strategy still hold up?
The standard technique is bootstrap resampling with replacement. You take your list of trades, shuffle them into new sequences, and recalculate the equity curve for each sequence. After 1,000 simulations you get a distribution of possible outcomes. If 80 percent of those simulations end in profit, your edge is probably real. If only 50 percent are profitable, the original result was likely driven by favorable trade ordering rather than a genuine edge.
I had a strategy that showed an impressive 2.4 profit factor over 300 trades. After 1,000 Monte Carlo runs, only 62 percent of simulations were profitable. The strategy was not bad, but the edge was much thinner than the headline numbers suggested. I would have deployed larger capital than was safe if I had looked only at the original backtest. This is the most common blind spot I see in backtest analysis: a strategy that crushes the historical test but falls apart when you randomize the trade order.
One caveat from my own use: Monte Carlo really wants at least 100 trades to produce trustworthy confidence intervals. Below that, the range of outcomes is too wide to act on. The simulation is still informative, but treat the results as directional rather than precise. I ran a strategy with only 45 trades through Monte Carlo and the 95 percent confidence interval stretched from a 12 percent loss to a 29 percent gain. That range tells you nothing actionable.
Recovery Factor and SQN: two underused quality metrics
Recovery Factor divides total net profit by maximum drawdown. It tells you how much you earned for every dollar you lost at the worst point. A Recovery Factor of 3.0 means you made three dollars for every dollar of peak-to-trough loss. Below 2.0 is weak. Above 5.0 is excellent.
What I like about Recovery Factor is that it penalizes strategies that take big drawdowns and take a long time to recover. Two strategies can have the same total return, but if one drew down 15 percent and the other drew down 5 percent, the second strategy has a much higher Recovery Factor and is safer to trade. Use the Recovery Factor calculator to check your strategy.
The System Quality Number (SQN) combines win rate, average trade return, and the standard deviation of returns into a single score. Developed by Dr. Van Tharp, it measures whether a strategy is tradeable. An SQN below 1.0 means the strategy is not worth trading. Between 1.0 and 2.0 is average. Above 3.0 is excellent. Above 5.0 approaches the level of the best professional traders.
When I first calculated SQN for a portfolio of strategies I was running, two out of five scored below 1.0. The win rates looked fine at 58 and 62 percent. But the SQN told me the variability was too high relative to the average trade. I dropped both strategies and the portfolio performance improved immediately. Try the SQN calculator on your own backtest data.
Rolling window analysis: spot strategy decay early
A strategy can have excellent full-sample metrics and still be degrading in real time. Rolling window analysis tracks metrics like Sharpe ratio, win rate, and Sortino ratio over a sliding window of your most recent trades. The standard window size is 20 trades.
I use rolling Sharpe the same way I check engine temperature in a car. When it starts trending down across several consecutive windows, something is wrong. In 2023, a momentum strategy I was testing showed a steady rolling Sharpe decline from 1.6 to 0.4 over 12 windows before the cumulative equity curve turned down. The rolling analysis gave me three months of warning that the strategy was breaking. I pulled the strategy before the drawdown hit. If I had been watching only the full-sample Sharpe, I would have stayed in and lost money.
Rolling metrics also reveal regime changes. A strategy that worked in trending markets may fall apart in sideways ranges. The rolling window catches these transitions early, while the full-sample average still looks acceptable. This is the single most practical feature for anyone actively running a strategy.
Watch for three patterns in your rolling charts: a persistent downward slope (the strategy is decaying), a sudden drop (regime change), and increased variance (the strategy is becoming unpredictable). Any of these is reason to pause trading and re-evaluate. I track rolling Sortino alongside rolling Sharpe because they sometimes diverge: a Sharpe goes down while Sortino holds steady, which means the volatility increase is on the upside and not a concern. When both trend down together, that is a stronger signal to stop and re-examine.
MFE/MAE analysis: optimize your exits with real data
Maximum Favorable Excursion (MFE) measures how far a trade went in your favor before it closed. Maximum Adverse Excursion (MAE) measures how far it went against you. Plotting these on a scatter chart reveals whether your exits are well calibrated.
I ran MFE/MAE on a breakout strategy that had a 58 percent win rate. The scatter plot showed something I had not noticed: most winning trades still had room to run further. The average winner was closed at 1.8R, but the average MFE on those same trades was 3.2R. I was leaving 44 percent of potential profit by taking profit too early. The exit logic was the bottleneck, not the entry.
MAE tells the other side of the story. If your losing trades consistently hit MAE values close to your stop loss, your stops are well placed. If MAE values cluster well below the stop level, your stops may be too wide. I tightened a stop by 15 percent after MAE analysis on a swing trading strategy and the win rate stayed the same while the average loss dropped by 12 percent. That was a pure improvement in risk management. Use the MFE/MAE analysis tool to examine your own trades. Pair the findings with the Strategy Optimizer to adjust entry and exit rules based on real trade data.
Together, MFE and MAE give you a data-driven way to set stops and targets instead of guessing. They are the difference between hoping your exits are right and knowing they are.
Kelly Criterion: sizing your bets correctly
The Kelly Criterion tells you what percentage of your capital to risk on each trade based on your strategy's win rate and average risk-reward ratio. The formula is simple: Kelly % = W minus ( (1 minus W) / R ), where W is the win rate and R is the ratio of average win to average loss.
Here is where it gets uncomfortable. Most retail strategies have a Kelly fraction below 10 percent. A strategy with 55 percent win rate and a 1.5 risk-reward ratio gives a Kelly of 18 percent. But that assumes your win rate and risk-reward are stable, which they rarely are. Most traders use half-Kelly or quarter-Kelly to account for estimation error.
I calculated Kelly on a strategy I was running and the result was 6 percent. That meant I should risk 6 percent of capital per trade. I had been risking 2 percent. I was underbetting the strategy, leaving returns on the table. On the other side, I have seen strategies where Kelly came back at 40 percent. That high number was a red flag: it usually means the sample is too small or the strategy is curve-fitted.
Your results may differ significantly from historical Kelly estimates because future markets are different from past ones. Treat Kelly as a upper bound, not a target. Half-Kelly is the standard conservative approach and has saved me from overbetting several marginal strategies.
Get all of it in one click with Pineify
Here is the problem with everything described above: calculating these metrics by hand takes hours and requires Python or Excel expertise. You need the raw CSV, a statistics library, and the time to verify each formula. Most traders do not have that time, so they skip the analysis and trade based on win rate alone. That is how good-looking backtests lose real money.
Pineify Backtest Deep Report does all of it in your browser. You export the CSV from TradingView, upload it to our tool, and get back a 16-KPI report covering every metric described above: Sharpe, Sortino, Calmar, SQN, Recovery Factor, Ulcer Index, UPI (Martin Ratio), VaR and CVaR, Skewness, Kurtosis, plus Monte Carlo simulation (1,000 bootstrap runs), rolling window analysis with three metrics, returns distribution histogram with normal curve overlay, MFE/MAE scatter with color-coded trades, heatmaps for monthly, weekly, daily, and time-of-day returns, Kelly Criterion, and an AI strategy verdict that scores your strategy from 0 to 100.
Everything runs client-side. Your CSV never leaves your device. There is no account required for the basic report. You can export the full analysis to an 8-sheet Excel workbook in one click. No Python, no R, no math degree needed.
Two honest boundaries: Pineify is not a backtesting engine. You run your strategy in TradingView and bring the results here for deep analysis. It is also not a trading journal. We do not sync with brokers or track trade psychology. But for reading the statistical truth of a backtest, it is faster and more thorough than any manual setup I have used. Open the CSV, drop it in, and the full report appears in under a second.
Start with the Backtest Deep Report or jump directly to a specific metric:
FAQ
Analyze your TradingView backtest in seconds
Upload your TradingView CSV and get Sharpe, Sortino, Monte Carlo, rolling analysis, and 16 more KPIs. All in your browser, nothing leaves your device.
Analyze My Backtest Now