Frequently Asked Questions

Getting Started

How many strategies can I import?

Limits depend on your plan — see your Subscription page for current usage and limits.

What file formats are supported?

AlgoChef accepts CSV, TXT (with various delimiters), TradeStation/Multicharts RINA XML, Excel (XLSX), and HTML report files — and you can import several files at once. See the Import Guide for details.

What's the minimum data needed?

At minimum, you need entry dates, exit dates, and P&L (in dollars or points) for each trade. More fields unlock more metrics.

How many trades do I need for reliable analysis?

The more trades, the more reliable the analysis. AlgoChef requires a minimum number of trades for scoring — below it, the Profitability, Risk, Confidence, and CSI scores aren't calculated and show "Not enough data". For the most robust statistical results, aim for as many trades as possible. The platform will guide you if your trade count is too low for specific features. See Why Strategies Fail — Insufficient Sample Size for more context on why this matters.

Does AlgoChef work with NinjaTrader or MetaTrader?

AlgoChef accepts CSV exports from any platform, including NinjaTrader and MetaTrader. Export your trade history as CSV from your platform and use the import wizard to map your columns. See the CSV Format Reference for details.

How do I report a bug or request a feature?

If you're logged in, use the Bug Report and Feature Request options available inside the app. For general questions, use the contact form on our website.

What is the difference between backtesting and paper trading?

Backtesting replays historical data through your strategy rules to measure how it would have performed. Paper trading runs the strategy forward in real-time on live market data without risking capital. Both are useful, but neither alone proves a strategy is robust. AlgoChef goes further by applying statistical validation — scoring your backtest across profitability, risk, confidence, and health dimensions — so you can quantify how trustworthy those backtest results really are before you paper trade or go live. See Why Strategies Fail in Live Trading for the most common failure modes.

Scoring

What makes AlgoChef's scoring system different?

Most platforms show you a handful of standalone metrics — Sharpe Ratio, net profit, max drawdown — and leave you to interpret them yourself. AlgoChef gives every strategy five proprietary scores — Profitability, Risk, Confidence, CSI (the overall quality verdict), and Health, which stands on its own and tracks whether the strategy is degrading. You get an institutional-grade snapshot of your strategy's quality in seconds, covering profitability, capital protection, statistical reliability, and ongoing degradation monitoring — all at once.

Why is my CSI score 0 or very low?

A CSI of 0 (shown as "Not enough data" in gray) means it hasn't been calculated — almost always because the strategy has fewer than the minimum number of trades. Import more trade history.

A very low CSI often means a circuit breaker fired. Circuit breakers force a Failed verdict when a fundamental problem is detected:

• Excessive drawdown — Strategy has too much capital risk
• No real edge — Strategy is not generating positive returns over its history
• Unrealistic win rate — Extremely low or high win rates may indicate a data issue
• Too good to be true — Implausibly high returns or risk-adjusted returns, or an equity curve too smooth to be real, often point to look-ahead bias or curve-fitting

The Quality Report names the breaker that fired. See Why Strategies Fail for a full explanation of the most common validation problems.

Why is my score low when my metrics look fine?

The CSI is a proprietary overall quality verdict, not an average of the metrics on your dashboard. It's built to reward strategies that are strong — not merely acceptable — where it matters most for surviving live trading, so a strategy where every visible metric looks acceptable can still land in the middle of the range. The related Profitability, Risk, and Confidence scores and the Quality Report's warnings often show where a strategy is struggling. See CSI — How is the score calculated? for more detail.

Why is my Confidence Score 0?

A 0 means the score hasn't been calculated: AlgoChef needs a minimum number of trades before it scores a strategy. Below that, Profitability, Risk, Confidence, and CSI all show "Not enough data". This is the most common reason a strategy has no CSI yet.

Can I improve my scores by adjusting settings?

Scores are based on your actual trade data. The only way to improve them is to improve the strategy itself (better entries, exits, risk management) or to provide more trade data for better statistical reliability.

What's the difference between CSI and Health Score?

CSI evaluates absolute quality — is this a good strategy? Health Score evaluates recent performance — is this strategy getting worse? A strategy can have a high CSI (historically good) but a low Health Score (recently degrading).

What is a good Sharpe Ratio for a trading strategy?

A Sharpe Ratio above 1.0 is generally considered acceptable, above 1.5 is good, and above 2.0 is excellent — but context matters. A high Sharpe alone doesn't guarantee robustness. AlgoChef's scoring system evaluates dozens of metrics across independent dimensions, so you're never relying on a single number.

What is a good Profit Factor?

Profit Factor (gross profit ÷ gross loss) above 1.0 means the strategy is profitable. Above 1.5 is solid, and above 2.0 is strong. However, a high Profit Factor on a small sample can be misleading. AlgoChef's Confidence Score measures whether your results are statistically trustworthy or just noise.

How many trades do I need to backtest a strategy reliably?

More is always better. With fewer than 30 trades, most statistical measures are unreliable. AlgoChef won't score a strategy below a minimum number of trades — the scores show "Not enough data" rather than a misleading number — and the Confidence Score tells you whether there's enough consistent evidence to trust the results once they are scored.

How do I know if my strategy is curve-fitted?

Curve-fitting (overfitting) means your strategy is tuned to past data rather than capturing a genuine edge. Warning signs include exceptional backtest results that don't hold up out-of-sample. AlgoChef's dedicated check is Overfit Likelihood, which looks at the structure of your trade history for the fingerprints hindsight fitting leaves behind. Monte Carlo simulation stress-tests your strategy under randomized conditions, and the IS/OOS validation compares in-sample performance against out-of-sample results — a large gap is a red flag for overfitting.

Strategy Validation

What is In-Sample vs Out-of-Sample testing?

In-Sample (IS) data is the portion of your trade history used to develop or optimize a strategy. Out-of-Sample (OOS) data is a held-back portion the strategy has never "seen." Comparing IS and OOS performance reveals whether your strategy captures a real edge or is just memorizing past data. AlgoChef automatically splits your trades and measures the divergence as part of the Health Score and Monte Carlo IS/OOS validation.

What is walk-forward testing?

Walk-forward testing repeatedly optimizes a strategy on a rolling in-sample window, then validates it on the next out-of-sample window. This process simulates how a strategy would perform if re-optimized periodically in real trading. It's one of the strongest defenses against curve-fitting because every validation segment is truly unseen data.

What is Monte Carlo simulation and why does it matter?

Monte Carlo simulation generates thousands of randomized variations of your trade sequence to see how results change under different conditions. It answers questions like "What's the worst drawdown I should realistically expect?" and "How likely am I to hit my profit target?" AlgoChef's Monte Carlo engine offers five simulation methods and recalculates all scoring dimensions for each run — giving you confidence intervals across 20+ metrics, not just a simple equity fan chart.

What is the difference between optimization and curve-fitting?

Optimization is the legitimate process of tuning strategy parameters to improve performance. Curve-fitting is optimization taken too far — the parameters become tailored to historical noise rather than a repeatable edge. The line between them is statistical: if performance collapses on unseen data, you've curve-fitted. AlgoChef helps you detect this through Overfit Likelihood (its dedicated check for hindsight-fitted backtests), the Confidence Score (is there enough consistent evidence?) and Monte Carlo IS/OOS validation (which compares in-sample and out-of-sample results).

Features

How does Monte Carlo simulation work?

Monte Carlo runs thousands of randomized variations of your trade history to test robustness. Unlike most tools that only simulate net profit and max drawdown, AlgoChef recalculates all four scoring dimensions (Profitability, Risk, Confidence, and CSI) for each simulation run — so you see confidence intervals across 20+ metrics, not just a single equity fan chart. See Monte Carlo Simulation for the five methods.

Can I combine strategies in a portfolio?

Yes! The Portfolios Hub offers two ways to build (pick it yourself, or let AlgoChef build it: Fast, Thorough, and advanced options) — and Portfolio Studio then analyzes, sizes, and stress-tests what you built.

What's the difference between Quick Build and Stability Composition?

Quick Build (the Fast button) gives you a sensible portfolio in seconds from a few choices. Stability Composition (the Thorough button) runs a correlation-aware engine in the background and hands you a ranked list of robust candidates. Speed versus thoroughness. See Portfolios Hub for details.

What is Risk of Ruin?

The share of simulations in which your account breaks a ruin rule. For futures, ruin means the account goes below the margin needed to hold the position, plus a buffer (default 10%); you can also add a drop from the account's highest point (default 50%). For forex, stocks and crypto (and futures with no margin requirement set), ruin means the account going below a dollar floor (default $3,500) or falling 50% (adjustable) or more below its highest point. It depends heavily on how much capital you start with — more capital leaves more room above the margin line or floor, and the same strategy can look comfortable at $50K and fragile at $10K. These are simulated results, not predictions. Monte Carlo Capital Sizing and the Portfolio Robustness check both compute it.

Why are some strategies marked "Sample · read-only"?

They're demo strategies included with new accounts so you can explore every feature before importing your own data. You can analyze them but not edit or delete them. See Account & Settings.

How often should I check the Health Monitor?

For actively traded strategies, check weekly (or every 10 trades). The Health Monitor compares recent performance against historical baselines — regular checks help catch degradation early.

How do I share a strategy comparison?

From a Head-to-Head comparison, click "Share" to generate a public link. Recipients can view the comparison without an AlgoChef account.

Can I run Monte Carlo on a portfolio?

Yes — the Robustness tab in Portfolio Studio runs Monte Carlo resampling, walk-forward validation, and stress scenarios on the whole portfolio in one check. The standalone Monte Carlo pages work on individual strategies.

How does AlgoChef compare to Strategy Quant or Market System Analyzer?

AlgoChef is a validation platform, not a strategy builder. Strategy Quant X generates and optimizes strategies automatically — AlgoChef validates whether those strategies are worth trading. Many traders use both together. Market System Analyzer focuses primarily on position sizing and Monte Carlo — AlgoChef covers the full validation workflow including scoring, portfolio construction, and health monitoring. See the full comparison page for details.

Data & Privacy

Is my trade data secure?

Your data is stored securely with row-level security — only you can access your strategies.

Can I delete my data?

You can delete individual strategies and portfolios anytime from their libraries. For full account deletion, contact support@algochef.app or see the Privacy Policy.

Does AlgoChef execute trades?

No. AlgoChef is a pure analysis platform. It does not connect to brokers, execute trades, or manage money. All trading decisions are yours.

Pricing & Trial

Is there a free trial?

Yes — AlgoChef offers a 14-day free trial on any paid plan. A card is required to start, nothing is charged until day 14, and you can cancel any time before then. Your trial includes everything on the plan you pick — and only that plan. See the pricing page for trial limits and plan details.

What happens when my trial ends?

Your data is preserved — nothing is deleted when a trial ends. AlgoChef is currently in invite-only beta and paid plans are not open for purchase yet. When they open you will be able to upgrade from the Subscription page and pick up exactly where you left off.

Troubleshooting

My import failed — what do I check?

• Verify the file format (CSV, TXT, or TradeStation/Multicharts RINA XML)
• Make sure date columns are in a recognizable format
• Ensure there's at least one P&L column (dollars or points)
• Check for special characters or encoding issues in the file

Metrics show N/A — why?

Some metrics require specific data:

• MAE/MFE require MAE, MFE value in dollars
• ATR metrics require ATR data in the import
• Profit Consistency requires 12+ months of data
• Stability metrics require a minimum number of trades

The scores don't match what I expected

AlgoChef's scores are designed to be conservative. It's hard to hit the Excellent tier, and even harder to score in the 90s. A strategy with great returns but poor risk management won't score well. A strategy with great stats on a short trade history won't score well either. That's by design. Most strategies score lower than traders expect because AlgoChef accounts for all trades including outliers and uses robust statistical methods.