Backtest glossary: report terms in plain English · Tradetron
🌐 USA ▾
Tradetron India USA
Book a demo Start free
Tradetron · Backtesting · Glossary

Backtest glossary: the report's terms, in plain words.

79 terms you meet in a Tradetron backtest report, in plain words: from CAGR, maximum drawdown and the Sharpe ratio to the probabilistic Sharpe ratio, Monte Carlo and walk-forward checks, costs such as slippage and STT, and option terms such as ATM, straddle and iron condor. Each term has its own link.

Run a backtest → See the data coverage
A to Z

All 79 terms.

01 · Glossary

Returns and headline numbers

#Capital required
The capital the engine reserved to run the strategy: the worst-case margin it needed at any point in the test. ROI, CAGR and drawdown percentages are all measured against this number.
#P&L (net and gross)
Total profit or loss for the period. Net is after brokerage, slippage and statutory charges, and is the number to judge a strategy on. Gross is before costs. In the report: P&L
#ROI
Return on investment: total P&L as a percentage of capital required.
#CAGR
Compound annual growth rate: the steady yearly return that would give the same end result over the same period. It lets you compare tests of different lengths.
#Sharpe ratio
Return earned per unit of volatility, on a yearly basis. A rough guide: below 1 is weak, 1 to 2 is decent, above 2 is strong. In the report: Sharpe
#Sortino ratio
Like the Sharpe ratio, but only falls count as risk. It is fairer to a strategy whose swings are mostly upward. In the report: Sortino
#Calmar ratio
CAGR divided by the maximum drawdown: yearly return earned for each unit of worst-case fall. Above 1 is good. In the report: Calmar
#Maximum drawdown
The worst fall from a peak to a low in the equity curve, as a percentage of capital required. The report also shows the rupee amount and the fall measured against the running equity peak (capital plus profits so far). In the report: Max DD · capital
#Win rate
The share of sessions that booked a result and finished positive (or of closed trades, in the trade card). Sessions that only opened a position are left out. A high win rate alone proves nothing: read it with expectancy and with average win against average loss.
#Expectancy
The average P&L of one trade once wins and losses are mixed together. If it is not positive, the strategy loses money over many trades however often it wins. In the report: Expectancy / trade
#Days traded
Sessions where the entry condition actually fired, out of all trading days in the tested period.
#Winning and losing streaks
The longest run of winning days and of losing days in a row. Expect a losing streak like it in live trading, and size your capital so you can sit through it. In the report: Streaks (trade days)
02 · Glossary

Verdict and robustness checks

#Grade
An overall rating from A to F that combines profit, risk, drawdown, consistency and how the strategy did across different periods. You can grade the gross (before costs) or net (after costs) result. The line next to it names the main reason for the grade.
#Probabilistic Sharpe ratio
How confident you can be that the strategy's Sharpe ratio is really above zero, given the number of trades, the shape of the profits and losses, and extreme days. 95% or more is strong evidence; below about 80% the result could be luck. In the report: Prob. Sharpe
#Walk-forward test
The history is split in two. The earlier part is where the strategy was tuned; the later part checks whether it still works on data it had not seen. "Holds up" means it kept working, "degrades" means it weakened, and "worse" means it is unlikely to survive live. In the report: Walk-forward
#Exit attribution
Shows whether the profit comes mainly from the exits you planned, or from exits forced on the strategy by a stop loss, a target, expiry, a roll or the end of the test.
#Worst rolling 12 months
The worst 252-trading-day stretch (about one year) in the test: what someone who started at the worst possible time would have gone through. In the report: Worst rolling 12M
#Concentration
How much of the total profit came from the best 3 months. A higher figure means the strategy depends on a few good months. Above 60%, check whether those months could happen again.
#Stale-quote fills
Fills priced off the last price that printed, not a price from that exact minute. Thinly traded strikes do not trade every minute, so the engine carries the last price forward for up to 30 minutes, as a live deployment does. A high count means the strategy trades contracts that barely trade, and those fill prices are the least reliable numbers in the report.
#Strikes filtered
Candidate strikes the backtest skipped during strike selection because their only price was too old to trust. Only the backtest applies this filter; a live deployment can still pick such a strike.
03 · Glossary

Risk and the spread of daily results

#Skew
How lopsided daily results are. Negative skew means many small wins and occasional big losses, which is typical of option selling. Positive skew is the reverse.
#Excess kurtosis
A measure of fat tails. Above about 3, extreme days happen far more often than a normal bell curve predicts, so the past maximum drawdown understates what is possible.
#Tail ratio
The size of a big up day compared with a big down day (the 95th against the 5th percentile). Above 1, the good tail outweighs the bad.
#Omega ratio
Gains divided by losses, each weighted by how likely it is, across all daily results. Above 1, the strategy made more than it lost. In the report: Omega
#Value at risk (VaR 5%)
On the worst 5% of days, you lose at least this much, as a percentage of capital. In the report: VaR 5%
#Conditional value at risk (CVaR 5%)
The average loss on those worst 5% of days. It is closer to what a really bad day feels like than VaR. In the report: CVaR 5%
#Best and worst day
The largest one-day gain and one-day loss, as a percentage of capital. In the report: Best / worst day
04 · Glossary

Monte Carlo

#Monte Carlo simulation
The report reshuffles the test's own trade results thousands of times (a bootstrap) to build many possible equity paths. It shows how much of the result depends on the luck of the order trades came in.
#Probability profitable
The share of the reshuffled paths that end in profit. In the report: Prob. profitable
#Terminal p5, median and p95
The final P&L of the reshuffled paths at the 5th percentile (an unlucky but plausible outcome), the median (the typical outcome) and the 95th percentile (a lucky outcome).
#Max drawdown p95
95% of the reshuffled paths kept their maximum drawdown inside this figure. Size your capital for this number, not for the single drawdown in the test. In the report: Max-DD p95
05 · Glossary

Benchmark comparison

#Alpha
Yearly return above what the strategy's exposure to the benchmark index would explain: the part that comes from the strategy itself. In the report: Alpha (ann.)
#Beta
How much the strategy moves when the benchmark moves. 0 means it does not follow the market; 1 means it moves one for one with the index.
#Correlation
How closely the strategy and the benchmark move together, from -1 to +1. Near 0 means the strategy adds something the index does not.
#Up capture and down capture
Up capture is the share of the benchmark's rises the strategy takes part in; higher is better. Down capture is the share of the benchmark's falls it suffers; lower is better. In the report: Up capture / Down capture
#Benchmark ROI
The buy-and-hold return of the benchmark over the same period: the bar the strategy has to clear.
06 · Glossary

Margin and recovery

#Peak margin
The highest margin the strategy needed at any single point in the test.
#Average margin
The average margin the strategy used across the test: roughly how much of your capital is working at a typical moment. In the report: Avg margin
#Return on peak margin
Profit compared with the margin actually needed to run the strategy. For example, ₹20,000 profit on ₹1,00,000 of margin is 20%.
#Ulcer index
How deep the losses went and how long they lasted, in one number. Lower is a calmer ride.
#Pain index
The average size of the loss the strategy was sitting in across the whole test. Lower means shorter and shallower losing periods.
#Recovery factor
Total profit divided by the biggest drawdown. Higher means the strategy earned back its worst fall more times over.
#Time to recover
The longest time, in days, the strategy took to climb back from a fall to a new high. In the report: Max time-to-recover
#Underwater
A strategy is underwater while it is below its previous high. The report shows the typical number of days it stayed there, and whether the test ends while it is still below its high. In the report: Median underwater / Currently underwater
07 · Glossary

Trades, sets and fills

#Trade (round trip)
One closed position: an entry matched to its exit. In the report: Trades
#Profit factor
The sum of all winning trades divided by the sum of all losing trades. Above 1.5 is healthy; just above 1 is fragile.
#Average win and average loss
The average P&L of winning trades and of losing trades. Read them with the win rate: a high win rate with a large average loss can still lose money. In the report: Avg win / Avg loss
#Average hold
The average time a trade stays open. In the report: Avg hold
#Fill
One executed order leg. A round trip is 2 fills and a rolled straddle is 4. The report counts entries and exits by the month each fill happened in.
#Lots traded
Total lots across all entry and exit fills: the turnover your broker charges on. The monthly "Lots" column counts only lots opened that month.
#Set
One entry condition of a strategy with its legs and its exit. A strategy can have several sets; the report shows how each one did.
#Diversification benefit
How much smaller the combined maximum drawdown is than the sum of each set's own maximum drawdown. 0% means the sets lose together; a high figure means one set's bad days are another's good days.
#Most correlated pair
The two sets whose daily P&L moves most alike. Above 0.7 they are largely the same trade twice, using twice the margin.
08 · Glossary

Costs

#Total costs
Brokerage, slippage and statutory charges, applied to every fill. The report's cost lab lets you change these assumptions and see the effect.
#Slippage
The gap between the price you expected and the price you got. The data has no bid/ask, so the backtest adds the slippage you set to each fill to stand in for the spread and price movement.
#Brokerage
The fee your broker charges per order. You set it in the backtest's cost profile to match your broker.
#STT
Securities Transaction Tax, charged by the government on trades on Indian exchanges. The rate depends on the instrument and on whether you buy or sell.
#Statutory charges
Charges set by law or by the exchange rather than by your broker: STT, exchange transaction charges, the SEBI fee, GST and stamp duty.
09 · Glossary

Data and how the engine runs

#1-minute bar (OHLC)
One minute of trading summed up as four prices: open, high, low and close. Backtests replay these bars, not individual ticks.
#Open interest (OI)
The number of futures or option contracts still open at a point in time. The India F&O data carries it for every minute.
#Exit on bar close
Stop losses and targets are checked on each 1-minute bar's close, not on every tick inside it. In a fast minute the exit can be past the level you set.
#Greeks
Delta, gamma, theta and vega: how an option's price responds to the underlying's price, to time and to volatility. The data has no stored Greeks, so the engine computes them from the option and underlying prices.
#Data lag
Backtests on NSE, NFO and BFO end a set number of days before today, following SEBI's norms on price data used for education. MCX, crypto and US backtests can end yesterday. The data coverage page shows today's latest end dates.
#Intraday
A run type where every position is closed the same day.
#Positional
A run type where positions can be held overnight and across days, for example until expiry.
#Live Offline
Running a strategy on live prices with every signal shown and no order placed. The step after a backtest and before trading with money.
10 · Glossary

Options and strategy terms

#Lot size
The fixed number of units in one futures or options contract, set by the exchange. Quantities in a strategy are in lots, and the exchange changes lot sizes from time to time.
#Expiry
The date a futures or options contract ends. Index options have weekly or monthly expiries; a strategy picks one, such as the current week.
#Strike price
The price at which an option lets you buy (a call) or sell (a put) the underlying.
#Premium
The price of an option. A seller receives it; a buyer pays it.
#ATM (at the money)
The strike closest to the underlying's current price.
#ITM (in the money)
An option that would be worth something if it expired now: a call with a strike below the current price, or a put with a strike above it.
#OTM (out of the money)
An option that would expire worthless at the current price: a call with a strike above it, or a put with a strike below it.
#Straddle
A call and a put at the same strike and expiry, usually ATM. Selling one profits if the price stays near the strike; buying one profits from a big move either way.
#Strangle
Like a straddle, but the call and the put are at different OTM strikes, so the price has a wider range before either leg is in the money.
#Iron condor
A sold strangle with a further-OTM call and put bought as protection. The bought legs cap the loss on either side.
#Stop loss
A rule that exits a position once its loss reaches a set level, in points, percent or rupees.
#Target
A rule that exits a position once its profit reaches a set level.
#Re-entry
Entering again after a stop loss or target has closed a position, when the entry condition is true again.

What history these numbers are computed on, and what a backtest cannot reproduce: backtest data coverage. A worked example: backtest Nifty options on 1-minute data.

Backtesting · Glossary

Read the numbers on your own strategy.

Build a strategy without code, backtest it, and every term above appears in your report.

Run a backtest →

Risk disclosure. For education and research only. Back-test results are hypothetical, simulated on historical data, and are not a claim, indication or assurance of returns or future performance. Not investment advice or a recommendation to buy or sell any security. Trading in securities and derivatives carries a substantial risk of loss. Tradetron is a software platform, not an investment adviser.