Glossary

Every term you meet in the app and on this site, in one or two plain sentences. Terms link to the ones they lean on.

53 terms

#

3D surface
A result drawn as a landscape over two settings, so broad hills and narrow spikes are visible at a glance.See also Plateau, Lone spike

B

Backtest
A run of a strategy over past prices, to see the trades it would have made. It shows what happened, not what will happen.See also Tick data, Strategy Tester, Overfitting

C

Cluster analysis
Sorts every setting into a few groups with similar results across the measures you pick.See also Pareto front, Plateau
Commission
The broker's fee per trade. In a MetaTrader 5 test it is part of each trade's result.See also Spread, Slippage
Complete optimisation
An optimisation that tests every combination of the settings you chose. In MetaTrader 5 it is called Slow complete algorithm. Every tool in the Report Analyzer works on it.See also Genetic optimisation, Optimisation
Correlation
How closely two strategies' results move together, from -1 to 1. Near 1, they win and lose on the same days.See also Correlation limit, Portfolio
Correlation limit
The highest correlation the Portfolio Optimizer allows between any two strategies in a mix. Any mix with a pair above it is dropped.See also Portfolio Optimizer

D

Daily loss limit
A prop firm rule: the most the account may lose within one day before it fails.See also Max loss, Prop firm
Data Bank
The runs you save in your account, kept by strategy, pair and file, ready to load into any module.See also Portfolio
Drawdown
A fall in the account from its last high to a later low.See also Max drawdown, Return/Drawdown

E

EA (Expert Advisor)
A strategy written as a program that MetaTrader 5 can run on its own, in a test or on a live account.See also MQL5, Magic number
.ex5
An EA compiled for MetaTrader 5 to run. It cannot be read as code.See also .mq5
Expected payoff
Net profit divided by the number of trades: what a trade made on average.See also Profit factor

G

Genetic optimisation
An optimisation that tests a sample of the combinations, steering towards the ones doing well. It is fast, but it leaves gaps between settings, so the tools that compare a setting with its neighbours show a notice instead of a result.See also Complete optimisation, Neighbour check

I

In-sample
The stretch of data a strategy was tuned on. A result here says little on its own, because the settings were chosen to fit it.See also Out-of-sample, Walk-forward (WFO)

L

Lone spike
A setting that does far better than every neighbour around it. Usually luck, and the first thing to drop.See also Plateau, Overfitting

M

Magic number
A number an EA stamps on its own trades, so trades from different strategies on one account can be told apart.See also EA (Expert Advisor), Reporting add-on
Max drawdown
The largest drawdown over the whole test.See also Drawdown, Return/Drawdown
Max loss
A prop firm rule: the most the account may lose in total, counted from the starting balance or from its high.See also Daily loss limit, Prop firm
MetaTrader 5
The trading platform many brokers offer for forex, CFDs and more. Strategywright reads its test results.See also Strategy Tester, EA (Expert Advisor)
Monte Carlo
Runs a strategy's own trades many times, drawn at random or in a new order, to show the spread of profits and drawdowns the same trades could have produced.See also Drawdown, Max drawdown
.mq5
An EA's source file, the code you can read and edit.See also .ex5, MQL5
MQL5
The language EAs are written in for MetaTrader 5.See also .mq5, .ex5

N

Neighbour check
Compares each top setting with the settings around it. A lone spike is flagged; a broad plateau is the one to trust.See also Plateau, Lone spike, Neighbour score
Neighbour score
How close a setting's neighbours come to its own result, from 0 to 1. 0.75 and up reads as a broad plateau, 0.5 to 0.75 as moderate, under 0.5 as a spike.See also Neighbour check

O

Optimisation
Running the same strategy many times, once for each combination of its settings, and recording every result.See also Genetic optimisation, Complete optimisation, Parameter (setting)
Optimisation export
The Strategy Tester's results table saved to a file. The Report Analyzer reads it as CSV, TSV, XML or ZIP.See also Optimisation
Out-of-sample
Data the strategy was not tuned on. A result here is the fairer test of the settings.See also In-sample, Walk-forward (WFO)
Overfitting
Tuning a strategy so closely to past prices that it fits their noise. It looks excellent in the backtest and has no edge after it.See also In-sample, Lone spike, Walk-forward (WFO)

P

Parameter (setting)
An input of a strategy you can change, such as the length of a moving average. Each optimisation row is one combination of them.See also Optimisation, Parameter sensitivity
Parameter sensitivity
How much the result moves as one setting changes, with the others left as they were. It shows which setting you cannot afford to get slightly wrong.See also Parameter (setting), Neighbour check
Pareto front
The settings no other setting beats on every measure you chose at once, such as more profit and less drawdown. You pick along it.See also Cluster analysis, Max drawdown
Plateau
A region of settings that all do about as well as each other. A pick from its middle still works if the market shifts a little.See also Lone spike, Neighbour check
Portfolio
Several strategies run together in one account, so one strategy's drawdowns can fall on another's good days.See also Portfolio Optimizer, Correlation
Portfolio Optimizer
Searches mixes of your strategies, drops any mix that breaks the correlation limit, and ranks the rest by Return/Drawdown, profit factor, win rate, Sharpe or Sortino. It never weights strategies.See also Portfolio, Return/Drawdown
Profit factor
Gross profit divided by gross loss. Above 1 the strategy made more than it lost.See also Win rate
Prop firm
A company that funds traders who pass its test account under fixed rules. The Prop Firm tool runs your trades through those rules.See also Daily loss limit, Max loss

R

Reporting add-on
A small MetaTrader 5 program that sends your closed trades and account balance to Strategywright through WebRequest. It has no code that places orders.See also WebRequest, Magic number
Return/Drawdown
Net profit divided by max drawdown. The Portfolio Optimizer ranks mixes by it unless you choose another measure.See also Max drawdown, Portfolio Optimizer
Rolling walk-forward (RWFO)
A walk-forward repeated period after period: tune, test the next period, move on by that period, tune again. You see how often the pick held up.See also Walk-forward (WFO)

S

Sharpe ratio
Average daily profit divided by how much daily profit varies, scaled to a year. Higher means steadier returns for the same profit.See also Sortino ratio
Slippage
The gap between the price an order asked for and the price it filled at.See also Spread
Sortino ratio
Like the Sharpe ratio, but only losing days count as variation, so upside swings are not punished.See also Sharpe ratio
Spread
The gap between the buy and the sell price at the same moment. Every trade pays it.See also Commission, Tick data
Strategy Tester
The part of MetaTrader 5 that runs backtests and optimisations of an EA. Its results table is what the Report Analyzer reads.See also Optimisation, Optimisation export
Swap
The fee or credit a broker charges for holding a trade overnight.See also Swap stress
Swap stress
Re-runs a strategy's results with its swaps multiplied, to see what higher overnight costs would do to it.See also Swap

T

Tick data
Every price change the broker recorded, not just one price per bar. A backtest on tick data fills orders closer to how they would have filled.See also Backtest, Spread
Trade Explorer
A table of every trade in a run, to sort and read one by one.See also What-If

W

Walk-forward (WFO)
Tune the settings on one stretch of the past, then test them on the period after it, which they never saw.See also Rolling walk-forward (RWFO), In-sample, Out-of-sample
WebRequest
MetaTrader 5's own way for a program to send data to a web address. It only reaches addresses you list in the terminal's settings.See also Reporting add-on
What-If
Changes a run's trades and shows the new results: a lower win rate, smaller wins, higher costs, or the worst losers removed.See also Commission, Win rate
Win rate
The share of trades that closed in profit. On its own it says nothing about how big the wins and losses were.See also Profit factor

Testing

11 terms
Backtest
A run of a strategy over past prices, to see the trades it would have made. It shows what happened, not what will happen.See also Tick data, Strategy Tester, Overfitting
Complete optimisation
An optimisation that tests every combination of the settings you chose. In MetaTrader 5 it is called Slow complete algorithm. Every tool in the Report Analyzer works on it.See also Genetic optimisation, Optimisation
Genetic optimisation
An optimisation that tests a sample of the combinations, steering towards the ones doing well. It is fast, but it leaves gaps between settings, so the tools that compare a setting with its neighbours show a notice instead of a result.See also Complete optimisation, Neighbour check
In-sample
The stretch of data a strategy was tuned on. A result here says little on its own, because the settings were chosen to fit it.See also Out-of-sample, Walk-forward (WFO)
Optimisation
Running the same strategy many times, once for each combination of its settings, and recording every result.See also Genetic optimisation, Complete optimisation, Parameter (setting)
Out-of-sample
Data the strategy was not tuned on. A result here is the fairer test of the settings.See also In-sample, Walk-forward (WFO)
Overfitting
Tuning a strategy so closely to past prices that it fits their noise. It looks excellent in the backtest and has no edge after it.See also In-sample, Lone spike, Walk-forward (WFO)
Parameter (setting)
An input of a strategy you can change, such as the length of a moving average. Each optimisation row is one combination of them.See also Optimisation, Parameter sensitivity
Rolling walk-forward (RWFO)
A walk-forward repeated period after period: tune, test the next period, move on by that period, tune again. You see how often the pick held up.See also Walk-forward (WFO)
Tick data
Every price change the broker recorded, not just one price per bar. A backtest on tick data fills orders closer to how they would have filled.See also Backtest, Spread
Walk-forward (WFO)
Tune the settings on one stretch of the past, then test them on the period after it, which they never saw.See also Rolling walk-forward (RWFO), In-sample, Out-of-sample

Results

14 terms
3D surface
A result drawn as a landscape over two settings, so broad hills and narrow spikes are visible at a glance.See also Plateau, Lone spike
Cluster analysis
Sorts every setting into a few groups with similar results across the measures you pick.See also Pareto front, Plateau
Expected payoff
Net profit divided by the number of trades: what a trade made on average.See also Profit factor
Lone spike
A setting that does far better than every neighbour around it. Usually luck, and the first thing to drop.See also Plateau, Overfitting
Neighbour check
Compares each top setting with the settings around it. A lone spike is flagged; a broad plateau is the one to trust.See also Plateau, Lone spike, Neighbour score
Neighbour score
How close a setting's neighbours come to its own result, from 0 to 1. 0.75 and up reads as a broad plateau, 0.5 to 0.75 as moderate, under 0.5 as a spike.See also Neighbour check
Parameter sensitivity
How much the result moves as one setting changes, with the others left as they were. It shows which setting you cannot afford to get slightly wrong.See also Parameter (setting), Neighbour check
Pareto front
The settings no other setting beats on every measure you chose at once, such as more profit and less drawdown. You pick along it.See also Cluster analysis, Max drawdown
Plateau
A region of settings that all do about as well as each other. A pick from its middle still works if the market shifts a little.See also Lone spike, Neighbour check
Profit factor
Gross profit divided by gross loss. Above 1 the strategy made more than it lost.See also Win rate
Sharpe ratio
Average daily profit divided by how much daily profit varies, scaled to a year. Higher means steadier returns for the same profit.See also Sortino ratio
Sortino ratio
Like the Sharpe ratio, but only losing days count as variation, so upside swings are not punished.See also Sharpe ratio
Trade Explorer
A table of every trade in a run, to sort and read one by one.See also What-If
Win rate
The share of trades that closed in profit. On its own it says nothing about how big the wins and losses were.See also Profit factor

Risk

13 terms
Commission
The broker's fee per trade. In a MetaTrader 5 test it is part of each trade's result.See also Spread, Slippage
Daily loss limit
A prop firm rule: the most the account may lose within one day before it fails.See also Max loss, Prop firm
Drawdown
A fall in the account from its last high to a later low.See also Max drawdown, Return/Drawdown
Max drawdown
The largest drawdown over the whole test.See also Drawdown, Return/Drawdown
Max loss
A prop firm rule: the most the account may lose in total, counted from the starting balance or from its high.See also Daily loss limit, Prop firm
Monte Carlo
Runs a strategy's own trades many times, drawn at random or in a new order, to show the spread of profits and drawdowns the same trades could have produced.See also Drawdown, Max drawdown
Prop firm
A company that funds traders who pass its test account under fixed rules. The Prop Firm tool runs your trades through those rules.See also Daily loss limit, Max loss
Return/Drawdown
Net profit divided by max drawdown. The Portfolio Optimizer ranks mixes by it unless you choose another measure.See also Max drawdown, Portfolio Optimizer
Slippage
The gap between the price an order asked for and the price it filled at.See also Spread
Spread
The gap between the buy and the sell price at the same moment. Every trade pays it.See also Commission, Tick data
Swap
The fee or credit a broker charges for holding a trade overnight.See also Swap stress
Swap stress
Re-runs a strategy's results with its swaps multiplied, to see what higher overnight costs would do to it.See also Swap
What-If
Changes a run's trades and shows the new results: a lower win rate, smaller wins, higher costs, or the worst losers removed.See also Commission, Win rate

Portfolio

5 terms
Correlation
How closely two strategies' results move together, from -1 to 1. Near 1, they win and lose on the same days.See also Correlation limit, Portfolio
Correlation limit
The highest correlation the Portfolio Optimizer allows between any two strategies in a mix. Any mix with a pair above it is dropped.See also Portfolio Optimizer
Data Bank
The runs you save in your account, kept by strategy, pair and file, ready to load into any module.See also Portfolio
Portfolio
Several strategies run together in one account, so one strategy's drawdowns can fall on another's good days.See also Portfolio Optimizer, Correlation
Portfolio Optimizer
Searches mixes of your strategies, drops any mix that breaks the correlation limit, and ranks the rest by Return/Drawdown, profit factor, win rate, Sharpe or Sortino. It never weights strategies.See also Portfolio, Return/Drawdown

MetaTrader 5

10 terms
EA (Expert Advisor)
A strategy written as a program that MetaTrader 5 can run on its own, in a test or on a live account.See also MQL5, Magic number
.ex5
An EA compiled for MetaTrader 5 to run. It cannot be read as code.See also .mq5
Magic number
A number an EA stamps on its own trades, so trades from different strategies on one account can be told apart.See also EA (Expert Advisor), Reporting add-on
MetaTrader 5
The trading platform many brokers offer for forex, CFDs and more. Strategywright reads its test results.See also Strategy Tester, EA (Expert Advisor)
.mq5
An EA's source file, the code you can read and edit.See also .ex5, MQL5
MQL5
The language EAs are written in for MetaTrader 5.See also .mq5, .ex5
Optimisation export
The Strategy Tester's results table saved to a file. The Report Analyzer reads it as CSV, TSV, XML or ZIP.See also Optimisation
Reporting add-on
A small MetaTrader 5 program that sends your closed trades and account balance to Strategywright through WebRequest. It has no code that places orders.See also WebRequest, Magic number
Strategy Tester
The part of MetaTrader 5 that runs backtests and optimisations of an EA. Its results table is what the Report Analyzer reads.See also Optimisation, Optimisation export
WebRequest
MetaTrader 5's own way for a program to send data to a web address. It only reaches addresses you list in the terminal's settings.See also Reporting add-on
TermMeaningSee also
3D surfaceResultsA result drawn as a landscape over two settings, so broad hills and narrow spikes are visible at a glance.Plateau, Lone spike
BacktestTestingA run of a strategy over past prices, to see the trades it would have made. It shows what happened, not what will happen.Tick data, Strategy Tester, Overfitting
Cluster analysisResultsSorts every setting into a few groups with similar results across the measures you pick.Pareto front, Plateau
CommissionRiskThe broker's fee per trade. In a MetaTrader 5 test it is part of each trade's result.Spread, Slippage
Complete optimisationTestingAn optimisation that tests every combination of the settings you chose. In MetaTrader 5 it is called Slow complete algorithm. Every tool in the Report Analyzer works on it.Genetic optimisation, Optimisation
CorrelationPortfolioHow closely two strategies' results move together, from -1 to 1. Near 1, they win and lose on the same days.Correlation limit, Portfolio
Correlation limitPortfolioThe highest correlation the Portfolio Optimizer allows between any two strategies in a mix. Any mix with a pair above it is dropped.Portfolio Optimizer
Daily loss limitRiskA prop firm rule: the most the account may lose within one day before it fails.Max loss, Prop firm
Data BankPortfolioThe runs you save in your account, kept by strategy, pair and file, ready to load into any module.Portfolio
DrawdownRiskA fall in the account from its last high to a later low.Max drawdown, Return/Drawdown
EA (Expert Advisor)MetaTrader 5A strategy written as a program that MetaTrader 5 can run on its own, in a test or on a live account.MQL5, Magic number
.ex5MetaTrader 5An EA compiled for MetaTrader 5 to run. It cannot be read as code..mq5
Expected payoffResultsNet profit divided by the number of trades: what a trade made on average.Profit factor
Genetic optimisationTestingAn optimisation that tests a sample of the combinations, steering towards the ones doing well. It is fast, but it leaves gaps between settings, so the tools that compare a setting with its neighbours show a notice instead of a result.Complete optimisation, Neighbour check
In-sampleTestingThe stretch of data a strategy was tuned on. A result here says little on its own, because the settings were chosen to fit it.Out-of-sample, Walk-forward (WFO)
Lone spikeResultsA setting that does far better than every neighbour around it. Usually luck, and the first thing to drop.Plateau, Overfitting
Magic numberMetaTrader 5A number an EA stamps on its own trades, so trades from different strategies on one account can be told apart.EA (Expert Advisor), Reporting add-on
Max drawdownRiskThe largest drawdown over the whole test.Drawdown, Return/Drawdown
Max lossRiskA prop firm rule: the most the account may lose in total, counted from the starting balance or from its high.Daily loss limit, Prop firm
MetaTrader 5MetaTrader 5The trading platform many brokers offer for forex, CFDs and more. Strategywright reads its test results.Strategy Tester, EA (Expert Advisor)
Monte CarloRiskRuns a strategy's own trades many times, drawn at random or in a new order, to show the spread of profits and drawdowns the same trades could have produced.Drawdown, Max drawdown
.mq5MetaTrader 5An EA's source file, the code you can read and edit..ex5, MQL5
MQL5MetaTrader 5The language EAs are written in for MetaTrader 5..mq5, .ex5
Neighbour checkResultsCompares each top setting with the settings around it. A lone spike is flagged; a broad plateau is the one to trust.Plateau, Lone spike, Neighbour score
Neighbour scoreResultsHow close a setting's neighbours come to its own result, from 0 to 1. 0.75 and up reads as a broad plateau, 0.5 to 0.75 as moderate, under 0.5 as a spike.Neighbour check
OptimisationTestingRunning the same strategy many times, once for each combination of its settings, and recording every result.Genetic optimisation, Complete optimisation, Parameter (setting)
Optimisation exportMetaTrader 5The Strategy Tester's results table saved to a file. The Report Analyzer reads it as CSV, TSV, XML or ZIP.Optimisation
Out-of-sampleTestingData the strategy was not tuned on. A result here is the fairer test of the settings.In-sample, Walk-forward (WFO)
OverfittingTestingTuning a strategy so closely to past prices that it fits their noise. It looks excellent in the backtest and has no edge after it.In-sample, Lone spike, Walk-forward (WFO)
Parameter (setting)TestingAn input of a strategy you can change, such as the length of a moving average. Each optimisation row is one combination of them.Optimisation, Parameter sensitivity
Parameter sensitivityResultsHow much the result moves as one setting changes, with the others left as they were. It shows which setting you cannot afford to get slightly wrong.Parameter (setting), Neighbour check
Pareto frontResultsThe settings no other setting beats on every measure you chose at once, such as more profit and less drawdown. You pick along it.Cluster analysis, Max drawdown
PlateauResultsA region of settings that all do about as well as each other. A pick from its middle still works if the market shifts a little.Lone spike, Neighbour check
PortfolioPortfolioSeveral strategies run together in one account, so one strategy's drawdowns can fall on another's good days.Portfolio Optimizer, Correlation
Portfolio OptimizerPortfolioSearches mixes of your strategies, drops any mix that breaks the correlation limit, and ranks the rest by Return/Drawdown, profit factor, win rate, Sharpe or Sortino. It never weights strategies.Portfolio, Return/Drawdown
Profit factorResultsGross profit divided by gross loss. Above 1 the strategy made more than it lost.Win rate
Prop firmRiskA company that funds traders who pass its test account under fixed rules. The Prop Firm tool runs your trades through those rules.Daily loss limit, Max loss
Reporting add-onMetaTrader 5A small MetaTrader 5 program that sends your closed trades and account balance to Strategywright through WebRequest. It has no code that places orders.WebRequest, Magic number
Return/DrawdownRiskNet profit divided by max drawdown. The Portfolio Optimizer ranks mixes by it unless you choose another measure.Max drawdown, Portfolio Optimizer
Rolling walk-forward (RWFO)TestingA walk-forward repeated period after period: tune, test the next period, move on by that period, tune again. You see how often the pick held up.Walk-forward (WFO)
Sharpe ratioResultsAverage daily profit divided by how much daily profit varies, scaled to a year. Higher means steadier returns for the same profit.Sortino ratio
SlippageRiskThe gap between the price an order asked for and the price it filled at.Spread
Sortino ratioResultsLike the Sharpe ratio, but only losing days count as variation, so upside swings are not punished.Sharpe ratio
SpreadRiskThe gap between the buy and the sell price at the same moment. Every trade pays it.Commission, Tick data
Strategy TesterMetaTrader 5The part of MetaTrader 5 that runs backtests and optimisations of an EA. Its results table is what the Report Analyzer reads.Optimisation, Optimisation export
SwapRiskThe fee or credit a broker charges for holding a trade overnight.Swap stress
Swap stressRiskRe-runs a strategy's results with its swaps multiplied, to see what higher overnight costs would do to it.Swap
Tick dataTestingEvery price change the broker recorded, not just one price per bar. A backtest on tick data fills orders closer to how they would have filled.Backtest, Spread
Trade ExplorerResultsA table of every trade in a run, to sort and read one by one.What-If
Walk-forward (WFO)TestingTune the settings on one stretch of the past, then test them on the period after it, which they never saw.Rolling walk-forward (RWFO), In-sample, Out-of-sample
WebRequestMetaTrader 5MetaTrader 5's own way for a program to send data to a web address. It only reaches addresses you list in the terminal's settings.Reporting add-on
What-IfRiskChanges a run's trades and shows the new results: a lower win rate, smaller wins, higher costs, or the worst losers removed.Commission, Win rate
Win rateResultsThe share of trades that closed in profit. On its own it says nothing about how big the wins and losses were.Profit factor