Workflow

A StrategyQuant X workflow in order: data, Edge Finder, Builder, MetaTrader 5

This is the order the modules were written to be used in. Stage 1 makes the symbol in SQX describe what your broker’s terminal trades. Stage 2 measures that symbol and configures a Build task. Stage 3 is StrategyQuant X as you already use it. Stage 4 runs the exported strategy on the files the research used. You can start at any stage.

Order of workFour stages from left to right: data, with SpecSmith and GapSmith; decide, with Edge Finder; build, which is StrategyQuant X with its Builder, retest, walk-forward matrix and export, and NameSmith; and trade in MetaTrader 5. Input modules feed the build stage. Edge Finder Deploy writes into the Build task.01 DATASpecSmith, GapSmithbroker profile matches MT5M1 gaps filled02 DECIDEEdge Finderside, hours, style, blocks03 BUILDStrategyQuant XBuilder, retestwalk-forward matrix, exportNameSmith names results04 TRADEMetaTrader 5same files, same weights,same calendarInputs for the BuilderInstitutional Indicators, NN Lab,Alternative Data, Newsblocks Edge Finder can measureDeploy writes sides, style and time range into the Build task.
01 · DATA

SpecSmith, GapSmith

Broker profile matches MT5; M1 gaps filled.

02 · DECIDE

Edge Finder

Side, hours, style and blocks; Deploy writes into the Build task.

03 · BUILD

StrategyQuant X

Builder, retest, walk-forward matrix, export, with NameSmith naming the results. Inputs: Institutional Indicators, NN Lab, Alternative Data, News.

04 · TRADE

MetaTrader 5

Same files, same weights, same calendar.

Fig. 1: The order of work. Box 3 is StrategyQuant X itself. Any module also works alone.
The path at a glance

Twelve steps, four stages

StepWhat you doToolWhereWhat it hands on
1aExport symbol specs, swaps, commissions and sessions from the broker’s terminalSpecSmith, script AWSpecExportMT5 Navigator, ScriptsTwo CSV files in Common\Files\AWSpecs
1bCompare with a broker profile, tick rows, applySpecSmithSQX Data Manager, SpecSmith tabA broker profile that matches the terminal, with a backup you can undo
1cCheck M1 history, fill holesGapSmithSQX Data Manager, Analyze data quality, Fix gapsA _patched symbol and an audit log; the original is untouched
2aSet symbol, timeframe, range and reserved barsSQX Data tabEdge FinderThe measuring sample
2bRun Autopilot, or each tabEdge FinderSQX left navigationMarket Bias, Time Windows, Style Fit and Block Edge results
2cApply to a Build taskEdge Finder, DeployDeploy tabSides, symmetry, style blocks, optional time range, read back from disk
3aRun the BuilderStrategyQuant XBuild taskCandidate strategies
3bRetest, walk-forward matrix and robustness checksStrategyQuant XSQXSurvivors
3cName the survivorsNameSmithDatabank actionNames like Breakout_SP500_H4_001
3dExport as an MT5 EAStrategyQuant XSQX exportEA, plus the AW indicators and data files it calls
4aPut the indicators and data files in the terminalInstaller, NN Lab deploy, feedsMT5 data folderSame history, model files, calendar
4bRun the EA in the Strategy Tester and compare with the SQX backtestMetaTrader 5MT5The parity check, before going live

Stage 1. Make the symbol match the terminal

A strategy is judged on the contract and the minutes the tester sees. Two things go wrong there. The contract can differ: SQX may hold a different tick size, point value, minimum stop distance, swap, commission or trading session than the broker’s terminal reports. And the minute data can have holes: tick-precision and M1 simulation build the price path from M1 bars, so a stop or limit inside a missing stretch fills at the next price that exists.

SpecSmith deals with the first. The AWSpecExport script is read-only; it writes two CSV files and changes nothing in the terminal. In the SpecSmith tab you compare, tick the rows you want, and apply. The values they replace are saved first, and Undo restores them and removes anything the apply created. It can also build a whole new broker profile from the terminal’s symbols. Spread is off by default, because the export holds one moment’s spread, and swap is never positive unless you allow it.

GapSmith deals with the second. It fills genuine gaps, never weekends, holidays or scheduled closures, up to the maximum gap length you set, and writes the result to a new symbol with a _patched suffix. Every filled minute is in an audit log. By default the filled bars are flat, at the previous close, with zero volume; an optional synthetic fill, marked and written to a separate _synth symbol, exists for genuine outages. They stop the simulation from jumping a gap and do not recover the prices that were lost.

SpecSmith copies what the terminal reports at the moment of export, and it does not guess the broker’s clock: MetaTrader gives only today’s UTC offset, so the timezone is yours to set. Its timezone check then tests your value against the weekly reopening of a currency pair and reports; it never changes your setting. GapSmith patches M1 history only.

Decide which symbol you will build on, the original or the patched one, and use the same one in stage 2 and stage 3. Test on both if a result looks too good around a patched stretch. See the data hygiene worked path.

Stage 2. Measure the symbol, then configure the Build task

On the Data tab choose the symbol, a timeframe and a range of several years, and add an out-of-sample range under Data range parts. A Build task uses the same control, so what you reserve here is reserved there. Press Autopilot, choose Current timeframe or Everything, and read the q grades and the Held out column before the means. Open Deploy, pick the project, check the side, symmetry, style and time range, and press Apply. The project must not be running.

Time Windows needs a data clock. Check the line at the top of its result before trusting any hour in it. Install the input modules first if you want them measured: Block Edge scans every block SQX has registered, including those added by Institutional Indicators, NN Lab, Alternative Data and News, and any you add yourself.

More: Edge Finder, Time Windows.

Stage 3. Build, retest and export in StrategyQuant X

Run the Build task as usual. Edge Finder and the input modules add a starting configuration and extra blocks; they do not change how SQX evaluates or ranks strategies. Keep the Build task’s out-of-sample range inside the window you reserved in Edge Finder and NN Lab, because a measurement or a model that saw those bars invalidates the validation. Then apply SQX’s own retest, walk-forward matrix and robustness checks to the survivors. Whether a strategy holds up is StrategyQuant’s verdict.

NameSmith names the survivors before you export, so the symbol, timeframe and style are in the name. Export as a MetaTrader 5 EA. Only MetaTrader 5 and PseudoCode are functional targets for these blocks.

  • Institutional Indicators: Wyckoff, order flow, structure and pattern blocks, with MT5 indicators.
  • NN Lab: trains MLP, convolutional, LSTM and CONVLSTM networks in SQX; one training run writes the SQX model and the matching MT5 model.
  • Alternative Data: eighteen non-price series with fixed publication dates.
  • News: blackout and timing blocks from your own MT5 calendar.

Stage 4. Run the same files in MetaTrader 5

The installer has already put the AW indicators in the terminals you ticked, and the feed services keep the alternative-data and news files current in Common\Files. Compile StrategyQuant’s own Sq* files once, and press Deploy to MetaTrader 5 in NN Lab if the strategy uses a model. Run the exported EA in the Strategy Tester first and compare its trades with the SQX backtest on the same history.

If the two disagree, work down the parity checklist: same instrument description, same history, same clock, same files, same indicators, closed bars, frozen versus live data, costs and fills, licence. See the parity checklist and why a backtest and MT5 disagree.

What stays outside

  • Costs, sizing and risk rules belong to the strategy and to your account. A funded account’s daily or trailing loss rule is not modelled by any module.
  • Robustness testing is StrategyQuant’s.
  • The history you trade on comes from your data provider and broker. The modules check and, where the rule allows, repair short gaps. They do not supply history.

Questions

Do I have to use every module?

No. Each works alone. The Suite includes all eight.

Why is data before Edge Finder?

Edge Finder measures whatever symbol and profile are on the Data tab. A symbol with holes, or a profile with a session that differs from the terminal’s, gives a measurement of a slightly different market. Edge Finder has a light patched-copy button of its own for the symbol on its Data tab.

Where do the walk-forward matrix and the retest sit?

After the Builder, in StrategyQuant X, on the survivors. Edge Finder works before the Builder.

Can I run Edge Finder after building?

Yes, on any symbol. Its Deploy step writes into a Build task, so it is most useful before a build.

Does the workflow need MetaTrader 5 at the start?

SpecSmith needs a terminal logged in to your broker, to export from. Edge Finder and GapSmith need only SQX.

Is any of this a promise of results?

No. Matching the backtest to the terminal removes some causes of disagreement. It does not make a strategy profitable, and no module predicts results.

Risk

Trading leveraged instruments carries a high risk of loss. Software output is not investment advice and says nothing about future results. Validate every strategy yourself before risking capital.