Tab Reference

EdgeProof's tabs live inside TradeStation's TradingApp, with two native tabs (Settings and Optimization) and several externally built tabs. A Docs tab at the front of the strip holds the app's own built-in reference material — open it any time, independent of the workflow below. Work through the rest left to right; the >> / << buttons on most tabs will take you to the next/previous step in the guided flow rather than simply moving one tab over — see the note at the end of this page for exactly where each arrow goes.

In this page

• Settings
• Strategy Code
• Optimization
• Trades Generator
• WFO Config
• Walk-Forward Run
• Robustness
• Candidates
• Validation
• Live Trading
• Docs
• A note on >> / << navigation

---

Settings

The Settings tab is native to TradeStation and it is the foundation of every session — everything downstream reads the parameters it writes to disk.

Set the Working Folder first. Nothing else on this tab, or any other tab, works without it.

Every pre-filled value on this tab is a starting default. Prefix/Suffix text, DaysBack, RV Percent, Commission/Slippage, the Metrics thresholds, MaxBarsBack, Trade Size, Genetic Optimization settings — all of it ships with a reasonable starting point, and all of it is yours to change. Review each field against your own strategy, instrument, and account before you rely on results, rather than assuming the shipped numbers were chosen for you.

Only the settings labeled with a * are required.

Work Environment

User Naming Convention

Strategy Selection

Data Settings

Costs

Optimization Settings

Metrics

The thresholds grid that filters which optimization results survive into the first, basic filtering — this is the gate described in Optimization mode below. Change the defaults as needed. The Basic Metrics count lets EdgeProof know how many of the first four metrics must individually pass. Enable Binomial toggles additional screening of the statistically-lucky results.

Trades Settings

Historical tick-data limits (a TradeStation platform limit, not an EdgeProof one). TradeStation only keeps about 6 months of tick-level history locally. Testing an expired futures contract older than 6 months at a Tick or Second Look-Inside-Bar resolution will fail with a data-acquisition error rather than complete — for deep historical testing, use Minute resolution instead. Look-Inside-Bar resolution also drives processing cost directly: a daily-chart backtest with 1-minute Look-Inside-Bar enabled evaluates roughly as many steps as a native 1-minute chart would, so optimizations can slow dramatically. Run your first parameter sweeps with Look-Inside-Bar set to None, and only turn it on for your shortlisted validation candidates.

Capital

Click Save Settings when done. If Working Folder, Strategy, or all optimization rows are incomplete, nothing is written and the status line reads:

Cannot save strategy contract yet. Please select Working Folder, Strategy, and complete at least one optimization row.

Once everything required is filled, the status line instead reads "Settings are ready. You can continue to Strategy Code." — click Save Settings and it confirms "Settings saved successfully." This writes StrategySetup.txt into the strategy folder. Any later edit to Working Folder, Strategy, or a grid cell immediately marks the contract dirty again and disables Save/Next until you save again.

Click >> to move to Strategy Code.

---

Strategy Code

Takes your raw EasyLanguage strategy and produces the optimization-ready version EdgeProof will run through the process.

Strategy Name field at the top mirrors the name from your saved Settings contract. It's editable here, and it's kept in sync with the same field on the Live Trading tab — change it on either and the other updates.

SOURCE CODE panel (left) — paste your raw strategy (copied straight from the TradeStation Development Environment). If the strategy already has a name in its code, pasting automatically re-reads the StrategyName(...) input in your source and may refresh the Strategy Name field to match it — if you're relying on the Prefix/Suffix name from Settings, paste your source before checking that the name field still reads what you expect.

RULES & FILTERS — five checkboxes, all unchecked by default:

Click Generate Code. The app parses your source, applies the checked filters, and writes the result to OPTIMIZED CODE panel, on the right. Status messages you may see:

• "Ready. Paste source code and click Generate." — initial state
• "Settings changed or not yet saved. Click Save Settings before generating code." — go back to Settings first
• "Warning: Please paste source code first."
• "Error: Parse failed. Check code format." — something in your pasted source doesn't parse; check for stray brackets or malformed Inputs/Vars sections
• "Success: Optimization strategy code generated and saved to Code\Optimization.txt."

The generated file is written to StrategyFolder\Code\Optimization.txt, regardless of what Prefix/Suffix or filter combination you used.

1. Copy Strategy Name.
2. In the TradeStation Development Environment: create a new strategy and paste the name in the wizard.
3. Back in EdgeProof, click Copy Code.
4. In the TradeStation Development Environment: paste the code and compile.
5. Return to EdgeProof and click >> to continue.

Clear wipes both panels and resets the status line to "Cleared. Ready for new strategy."

---

Optimization

One tab, three modes (defaults to Optimization):

1. Feasibility — (Mode 0); selected only by the user, if needed.
2. Optimization — (Mode 1); selected automatically by the process or by the user in the Test Mode dropdown.
3. Validation — (Mode 2); selected automatically by the process or by the user in the Test Mode dropdown.

Everything on the tab — button captions, visible panels and messages — changes with the mode.

Note: To avoid confusion — "the Optimization tab" is this whole tab; "Test Mode: Optimization" is one of its three modes.

The Inputs grid and values (Mode 1 and Mode 2) are already filled — carried over from your saved settings. All other settings are defaults, not fixed. Change Symbols and Intervals of DATA1 to DATA4, as needed, before running.

Symbols and data streams (all modes):

• DATA1 Symbols / DATA1 Intervals — the main data, where all Trades are executed; comma-separated, e.g. @ES,@NQ / 5,10.
• DATA2 — optional. Can be either:
  • Higher timeframes of the same Symbols as DATA1, in which case all symbols from DATA1 have to be here too.
  • VIX.X — both Symbol and Interval required.
• DATA3 / DATA4 — optional supporting streams. Each requires both Symbol and Interval filled, or both left blank — filling only one throws "DATA2 requires both Symbol and Interval." (same message pattern for DATA3/DATA4).

Test Mode: Feasibility

A quick single-pass sanity check before committing to a full sweep — seconds, not hours. It only uses the first symbol you've entered even if you've listed several; if you need to check more than one symbol, run Feasibility once per symbol or move straight to a full optimization.

There's no field validation gate on this mode beyond the optimization grid itself — an empty Symbol won't be caught with a friendly message, so fill Symbol and Interval before clicking. If no input row has Start, Stop and Step set, you'll see "No optimization ranges defined — set Start/Stop/Step for at least one input in the grid."

1. Set Test Mode to Feasibility.
2. Confirm Symbol/Interval and at least one Start/Stop/Step pair are filled.
3. Click Run Feasibility Test.

When it finishes: "Feasibility Test Complete! Net Profit: X Max Drawdown: Y", and a Pass (green) / Fail (red) indicator based on Net Profit > 1000 and Net Profit / |Max Drawdown| > 1.2. Nothing is written to disk in this mode — it's a look, not a record. Click Abort Feasibility Test to cancel; there's no pause/resume in this mode, only a hard cancel.

Test Mode: Optimization

The full parameter sweep — every combination (Exhaustive) or a genetic search, across every symbol/interval pair, filtered by the Metrics thresholds grid on Settings.

Mandatory before Start: Symbol and Interval filled for DATA1 through DATA4, as your strategy requires. Missing settings will show: "Fill all mandatory fields and Start again!"

1. Set Test Mode to Optimization.
2. Click Start Optimization.

The Optimization Progress grid fills in as tests complete. When the whole batch finishes, results are written to a CSV per symbol, then filtered through the Metrics thresholds. The surviving parameter sets are written to Selected\SelectedParams_<date>.csv.

Click Abort Optimization once, to pause execution. Pause happens after the current job completes, and the "Start Optimization" button relabels to "Continue Optimization". Results completed so far are kept. Click Continue Optimization whenever ready to execute the remaining tests. To cancel execution, click Abort Optimization twice.

---

Trades Generator

This is just a status grid. Trades generation is fired automatically, if the Generate Trades checkbox has been checked in the Optimization tab. Otherwise, click Generate Trades — this tab only shows the status of each Trades file.

"Complete" only means the file was written successfully — it isn't a quality signal on its own. There's no re-run button on this tab; if you need to regenerate, go back to the Optimization tab and run Generate Trades again.

Your data stays local. Every trades file EdgeProof generates is written as a plain, unencrypted .csv under StrategyFolder\Trades\YYYYMMDD\ — on your machine, not uploaded anywhere. Your strategy logic, parameter sweeps, and trade history stay under your control throughout the pipeline.

Before anything has run you'll see: "Waiting for Trades Generator to run…" Once the batch finishes, a green banner confirms: "✓ Generation complete — {N} trades files generated."

When every row shows Complete, you're ready for either:

• Candidates — for a closer look at individual parameter sets, or
• WFO Config — the primary next step, to build your walk-forward plan

Click >> to continue to WFO Config.

---

WFO Config

This is the stage where you commit to a Walk-Forward plan, based on the results shown, and hand it to Validation. Selecting a Symbol/Interval pair from the dropdown at the top immediately loads every matching candidate and builds the full scoring matrix.

The matrix has 7 columns (4 through 10 segments) and 6 rows (10% through 35% OOS), 42 cells total. Each cell scores a Walk-Forward simulation across every candidate for that segment count and OOS split, using a fast approximation that slices the trade files you already have (this is not the real segment-by-segment re-optimization — that happens during Validation).

Cell value is an efficiency ratio — OOS return per day divided by IS return per day:

• Blank / pink — at least one segment lost money out-of-sample, or there wasn't enough data to score. Never pick these.
• Orange/yellow — under 1.0: OOS underperforms IS. Marginal.
• Green, darker with higher value — 1.0 and above: OOS holding its own or outperforming IS on a per-day basis. From 2.0 up, the cell renders as solid dark green.

Click a scored cell to see the detail behind it: a ranked list of every candidate's out-of-sample performance for that configuration, and a per-segment breakdown showing which parameter set won each in-sample window and how it performed out-of-sample. This click also immediately saves the selection to disk (LockedWfoPlan.txt), so navigating away and back restores your last choice.

Click Add to WFO to commit the selected cell — this writes the ValidParams file that Test Mode: Validation will run against. On success, the message box confirms "Candidate added!" along with the run's details. If the file no longer exists for the winning candidate (e.g. the trades folder was regenerated since you loaded the matrix), you'll be asked to reselect: "Matrix is stale - its trades files no longer exist (regenerated?). Re-select the Symbol/Interval pair and click a matrix cell again."

To validate more than one configuration in the same session, repeat — select a cell, inspect it, click Add to WFO — one cell at a time for each Symbol/Interval or matrix combination you want to run.

Avoid a single green cell surrounded by blanks or orange — that's usually a sampling artifact, not a robust configuration. Prefer a cluster of similarly-scored green cells across neighboring Segment/OOS% values, and check that different segments in the detail view picked different winning parameter sets — if every segment picked the same winner, the strategy isn't really demonstrating adaptability.

Click >> to continue to the Optimization tab, Test Mode Validation, and click Start Validation to run it.

---

Walk-Forward Run (Optimization tab)

Test Mode: Validation

Runs the Real Walk-Forward Optimization plan you locked in on the WFO Config tab. This mode has no field validation of its own — Symbol, Interval, and every date window come from the ValidParams file WFO Config wrote. If that file isn't there yet: "Real WFO: ValidParams not found. Run WFO Config first."

1. Complete WFO Config first — the section on that tab explains how.
2. Set Test Mode to Validation.
3. Click Start Validation.

Click Abort Validation once, to pause execution. Pause happens after the current job completes, and the "Start Validation" button relabels to "Continue Validation". Results completed so far are kept. Click Continue Validation whenever ready to execute the remaining segments. To cancel execution, click Abort Validation twice.

From here the app runs unattended: for each segment in your WFO plan it optimizes on the in-sample window, automatically picks a winner, locks those parameters, and re-runs the out-of-sample window. It repeats this for every segment, stitches the out-of-sample trades together, and — if Real Validation was on when you built the plan — runs one final backtest over the held-out tail. If you've queued more than one WFO configuration, it moves straight on to the next once the current one finishes; you'll see "All queued Real WFO runs complete!" when there's nothing left.

There's no separate progress detail for this mode beyond the shared progress bar and the message log — the on-screen segment counter is cosmetic and shouldn't be read as an exact position indicator. Watch the message log for the real state.

Pause and resume works the same way here as described above for Optimization mode. The exact messages you'll see: pressing Abort Validation once shows "Pause requested - will pause after the current job completes. Press Abort again to cancel the current job immediately." Once paused: "PAUSED - {position}. Press Start to resume from this point." Press Abort Validation a second time while a pause is armed to hard-cancel instead — "Cancelled - the interrupted unit left no completion marker. Press Start to retry from the last completed step." If a job fails outright, the same resume mechanics apply: "JOB FAILED... Position preserved - press Start to retry from the failed step." Progress is tracked on disk, so this survives closing and reopening TradeStation entirely — not just the run.

---

Robustness

A visual check on how stable your optimization results are across nearby parameter values — separate from, and complementary to, the plateau logic that runs automatically during Validation.

Pick an Optimization File — one of the raw result CSVs your strategy folder accumulates from Test Mode: Optimization — from the dropdown, or Browse… to one directly. Open Selected File opens it outside the app if you want to inspect the raw numbers.

The main view is a 3D surface: two of your swept parameters on the horizontal axes, Net Profit on the vertical axis, colored green (high) through yellow/orange to red (low). Read it the same way you'd read the WFO Config matrix's color coding — a broad, smoothly-sloped green region means nearby parameter combinations behave similarly (a robust plateau you can trust), while a sharp, narrow spike surrounded by red is a warning sign: the optimizer found one lucky combination rather than a genuinely durable edge.

---

Candidates

Your analysis hub for individual trade files, reading whichever trades file the tab was opened with. There's no forward/back wizard button here; open it any time once trade files exist. This tab opens as an embedded view of EdgeProof's companion web app.

Tab strip: Trades · Equity Curve · Analytics · Deep Analysis · Time · Monte Carlo · Risk · WFA · Settings, plus an AI Assistant tab that only appears when running inside TradeStation.

• Trades — a sortable, filterable blotter (Direction filter, Profitable-only toggle, live trade count and total P&L, CSV export).
• Equity Curve — cumulative equity with a range toggle (1M/3M/6M/1Y/ALL), an IS/OOS split control, and automatic annotations: shaded drawdown/recovery regions, Best/Worst trade markers, and footer KPIs (Max Drawdown %, Annualized ROI, GHPR).
• Analytics — a KPI dashboard (Net Profit, Profit Factor, Win Rate, Expectancy, Sharpe, Sortino, Calmar, Max DD, Recovery Factor, and more) alongside Monthly Returns, Drawdown, Rolling Sharpe/Sortino, Rolling Win Rate, Trade Distribution, and Long-vs-Short charts, a Monthly Heatmap, a Duration-vs-PnL scatter, and By Direction / By Symbol breakdown tables.
• Deep Analysis — opens with the binomial "randomness" test (is this win rate better than a coin flip, given the trade count?), then — where your trade CSV includes the relevant columns — cost-drag, MAE/MFE, and bars-in-trade charts.
• Time — three views (Tables, Heatmap, Calendar) breaking performance down by hour of day, day of week, and month.
• Monte Carlo — five views; see below.
• Risk — an R-multiple risk breakdown (normalized to TradePnL ÷ |MAE|) and a slippage-sensitivity sweep showing your strategy's break-even ticks-per-side and a survival table across a range of slippage assumptions.
• WFA — exploratory walk-forward analysis on this one candidate's executed trades: a segment-by-segment "Test Results" view and a "Cluster Analysis" heatmap sweeping segment count × OOS%, both governed by a shared, editable PASS/FAIL rules popover. This is exploratory only — it does not feed the actual locked WFO plan, which is decided on the WFO Config tab.
• Settings — a read-only summary of this candidate's symbol, interval, date range, strategy, inputs, and trade size, pulled automatically from the per-candidate settings file EdgeProof writes alongside the trades CSV. Blank if no matching settings file was found.
• AI Assistant (TradeStation only) — a chat panel over your currently loaded trades: a row of preset questions plus a free-text box, some answered instantly from local computation, others via a live AI service. Only your typed question is sent to that service — your trade data and strategy logic stay on your machine.

Monte Carlo

Five views, each answering a different robustness question about the selected candidate.

Forensic Deep-Dive — set Iterations (default 5,000), Seed, Mode (Shuffle, reordering your actual trades with every P&L value unchanged, or Bootstrap, resampling trades with replacement), and a Capital figure for an inline Risk-of-Ruin readout. Results show the simulated median and 5th-percentile-worst outcomes alongside your actual result, plus two histograms — Final Equity and Max Drawdown distributions. A tight histogram means trade order barely matters; a wide one means the real result depended on favorable sequencing.

Fan & Streaks — an equity "fan chart": a cone of simulated equity paths at each trade index (5th/25th/50th/75th/95th percentile bands), with your actual historical equity overlaid — near the top of the cone is a lucky run, near the bottom unlucky, the middle is typical. Alongside it, a distribution of the longest losing streak seen across simulations. A Block bootstrap mode is available here for strategies where trades are serially correlated (a winning trade tends to follow another winning trade, or vice versa) — it resamples contiguous chunks instead of individual trades, which usually widens the fan when that correlation is real.

Probability Analysis — two cumulative-probability curves: what fraction of simulations achieved a given net profit or better, and what fraction experienced a given drawdown or worse. Use the drawdown curve to size positions, not the single historical worst case.

Batch Validation — reruns the shuffle test across 10 fixed random seeds for this one candidate and reports Original DD vs. simulated worst-case DD per seed, with a Pass/Fail (Lucky) status per seed — a robustness check on this specific parameter set, not a scan across your whole candidate pool.

Risk & Scaling — choose what to sweep (Capital ($) or Bet size (×)), a simulation Mode (Bootstrap / Block bootstrap / Shuffle), and click Analyze Risk. Produces a table of tiers with Risk of Ruin percentage, median PnL, median drawdown, and a worst-case (5th percentile) drawdown at each tier — color-coded by Risk of Ruin. A scatter chart alongside it plots final equity against max drawdown across every simulated scenario, with crosshairs at the median.

---

Validation

Shows the outcome of a completed walk-forward Validation run — how the strategy performed on data it was never optimized on, one segment at a time.

WFO Run dropdown lists completed runs (newest first). Selecting one populates two inner sub-tabs.

Simulated

The main comparison: your walk-forward result (WFO, blue) against a full-history-best baseline (Initial, red) restricted to the same out-of-sample windows so the two are comparable apples-to-apples. If no exact matching baseline can be found, the app falls back to the nearest candidate and flags it: "Exact param match not found. Arbitrary baseline used for comparison."

• Equity chart — the two curves overlaid.
• Drawdown chart and WFE-per-segment bar chart — side by side underneath.
• Metrics table — Net Profit, Max DD, Win Rate, Profit Factor, Avg Trade, Trades, each with WFO / Initial / Diff% columns. Diff% is colored so a negative divergence stands out.
• Per-segment table — OOS dates and the winning parameter values for each segment, plus a per-segment OOS PnL bar chart.

Reading it: WFO tracking or beating Initial is a strong result — the strategy generalizes. WFO clearly below Initial means real-world, walk-forward performance falls well short of what a full-history optimization implied — normal to some degree, concerning if the gap is large. If the app can't establish out-of-sample date windows for the baseline at all, it hides the Initial line rather than show a misleading comparison: "OOS windows not found (ValidParams missing?). Initial baseline hidden."

Real

Only appears if Real Validation was enabled when the plan was built. Shows a single equity curve and the same metrics table, this time for the one final holdout backtest run over the tail of history that was never touched by any optimization or walk-forward segment — the closest thing this app produces to an out-of-time test. No WFO-vs-Initial comparison here; it's one number, standing alone. Appears automatically once a Real Validation run completes.

Click >> to continue to Live Trading.

---

Live Trading

Produces a clean, production-ready version of your strategy — trade recording and every optimization scaffold stripped out, and your walk-forward winning parameters baked in as the defaults. This is the last tab; there's no >> forward from here.

Strategy Name field is synced with the same field on the Strategy Code tab.

SOURCE CODE panel (left) is a reference view only — click Load Strategy to pull the current Optimization.txt into it. Editing this panel does nothing to the generated output — Generate Code always re-reads Optimization.txt straight from disk, regardless of what's showing here. Use Load Strategy to eyeball what you're about to generate from, not to make changes.

Click Generate Code. It:

1. Reads Code\Optimization.txt fresh.
2. Detects which of the RTH / Day P&L / EOD Exit filters were baked into it (no checkboxes to set here — whatever was checked on the Strategy Code tab carries over automatically).
3. Strips all trade-recording and optimization-scaffold code.
4. Injects your walk-forward winning parameters as hardcoded default Input values — pulled automatically from the most recent Validation run's results.
5. Renames the strategy from ...__Opt to ...__Live, only if your generated name actually contains the text __Opt — which it won't with the shipped Prefix/Suffix defaults. If it doesn't match, the name carries over unchanged and you should rename it yourself before compiling.
6. Prepends a short diagnostic header to the output confirming how many header/input/variable/body lines were parsed — this is informational, not an error, even though it looks like one at a glance.

Output is saved to Code\4_Live.txt and shown in LIVE TRADING CODE on the right. If no walk-forward parameters were found to inject, the status line notes it: "…(No WFO params injected)." — worth double-checking before you compile if you expected specific values.

1. Review the diagnostic header at the top of the output.
2. Click Copy Code.
3. In the TradeStation Development Environment: create a new strategy, paste, compile.
4. Apply the compiled strategy to your live chart. This last step is entirely manual — EdgeProof doesn't attach strategies to charts for you.

Clear resets both panels to "Cleared. Ready for new strategy."

---

Docs

Reference material built into the app itself — the app's own documentation, available any time from the front of the tab strip, independent of where you are in the workflow.

---

A note on >> / << navigation

These buttons jump to a specific named tab, not simply "next tab over" — and a couple of them branch depending on which Test Mode is currently selected:

• Settings >> → Strategy Code
• Strategy Code >> → Optimization
• Optimization << → WFO Config if Test Mode is Validation, otherwise Strategy Code
• Optimization >> → Trades Generator if Test Mode is Optimization; Validation if Test Mode is Validation; in Feasibility mode it shows a reminder instead of navigating: "Use Test Mode = Optimization to continue to Trades Generator, or Test Mode = Validation to continue to Validation."
• Trades Generator >> → WFO Config
• WFO Config >> → back to the Optimization tab (Validation mode)
• Validation << → Optimization tab (Validation mode); >> → Live Trading
• Live Trading << → Validation; there is no >>

Candidates has no >>/<< at all — reach it by clicking its tab header whenever you want a closer look at a candidate, at any point after trade files exist.