Demo AvailableResearch
Performance & Risk Research
Drawdown analytics, performance visualization, historical strategy evaluation, exposure and scenario analysis.
Trading Terminal · Research & Portfolio
Backtest Factory
Run thousands of backtests across a grid of parameters, see which regions hold up, and stress-test results with Monte Carlo resampling.
In Development — Targeting December 2026
BTFBacktest FactoryNIFTY · SHORT STRADDLE · LOT 65Concept render · illustrative data
Strategy builderSYNTHETIC
SELL ATM CESELL ATM PE
Weekdays
Expiry
CostsALWAYS ON
Synthetic NIFTY 2025-10-01 → 2026-09-18 · 253 sessions · 5-min bars · weekly expiry Tue. Splits by date: train 60 / validation 20 / out-of-sample 20.
10,000/ 10,000 backtestsCOMPLETE1,643 jobs/s · Short Straddle
Entry × Stopmedian train ₹/trade
09:2009:2509:3009:4510:0010:1510:3011:0011:3012:001.11.151.21.251.31.41.51.61.82
Target × Stopmedian over entries
10%15%20%25%30%40%50%60%75%hold1.11.151.21.251.31.41.51.61.82
IV × VIX at entryselected · mean ₹/trade
<12.2≥12.2≥13.4≥14.4≥15.9<13.4≥13.4≥14.5≥15.5≥17.2
Weekday × Expirymedian train ₹/trade
MONTUEWEDTHUFRICurrent wkNext wk
Outlined = broad stable region · prefer it over the brightest single cell
Result10,000 HYPOTHETICAL BACKTESTS COMPLETE
Selected cell · defaultE 09:45 · S 1.5× · T 10%
Short Straddle · 09:45→15:15 · target 10% · stop 1.5× · next-wk · 1×65
Default selection = strongest 3×3 neighbourhood (not a recommendation). Click any cell.
- Trades
- 253
- Net P&L
- +₹1,57,896
- Expectancy
- +₹624
- Profit factor
- 1.78
- Max drawdown
- −₹30,818
- Sharpe / Sortino
- 3.21 / 3.85
- Avg MFE / MAE
- +2.1k / −1.5k
Equity · net after costs
Robustness · by date
| TRAIN | VALID. | OOS | |
|---|---|---|---|
| Trades | 151 | 51 | 51 |
| ₹/trade | +₹770 | +₹503 | +₹313 |
| Net | +116k | +26k | +16k |
| PF | 2.45 | 1.42 | 1.26 |
| Max DD | −13k | −23k | −29k |
Monte Carlo · 1,000 runs · ruin line −20% · ruined 14
Testing many combinations raises the chance a result holds only by luck — prefer broad stable regions.
HYPOTHETICAL · SYNTHETIC DATA. Capital basis ₹2,00,000. Simulated results do not indicate future returns. Not advice.
BTF · HYPOTHETICAL results on SYNTHETIC data · costs always on · research tool, not adviceSIMULATED DATA · NOT LIVE · IN DEVELOPMENT
Demo data · simulated feed. Illustrative only — not live market data and not investment advice. Try it: click rows, tabs and controls.
Demo AvailableA working demonstration exists and has been shown publicly. It is not yet available to users.
Overview
A single backtest is one roll of the dice: change the entry time or the stop slightly and the result can look completely different. Backtest Factory runs the same strategy across thousands of parameter combinations at once — entry time, stop, target, expiry and more — and shows the results as heatmaps, so you can see whether a result sits on a broad, stable region or on a lucky peak.
Results are split into training, validation and out-of-sample periods, costs and slippage are always included, and a Monte Carlo survival chamber reshuffles the trade sequence thousands of times to show the range of drawdowns the same trades could have produced. Everything is hypothetical research on historical data.
Demonstrated in the ORVINUS feature series
Prototype
From the Prototype
A screen from the working prototype shown in the ORVINUS demonstration series.
Capabilities
What It Is Designed to Do
Intended capabilities. The released feature may differ.
Parameter sweeps
Thousands of combinations of entry, exit, stop, target and filters tested in one run.
Stability heatmaps
Results by parameter pair, so broad stable regions stand out from isolated peaks.
Out-of-sample checks
Training, validation and out-of-sample periods reported side by side.
Monte Carlo survival
Thousands of reshuffled trade sequences with drawdown percentiles and risk-of-ruin.
Costs included
Brokerage, taxes and slippage applied to every simulated trade.
In practice
Workflow & Limits
Intended workflow
- Define a strategy and the parameter ranges to test.
- Run the factory and watch the grid fill.
- Look for broad, stable regions rather than the single best cell.
- Check out-of-sample results and the Monte Carlo range before trusting anything.
What to keep in mind
- Backtests are hypothetical results on historical data. They do not guarantee or indicate future performance, and testing many combinations increases the risk of finding results that hold only by chance.
- Historical data depth depends on the data available for each instrument.
- This page describes a feature in development. It is not available to use yet.
Related
Works Well With
Demo AvailableResearch
Bot Swarm
Many small paper-trading bots, each watching one specific setup — when one fires, you see exactly which conditions it checked and why.
In DevelopmentResearch
Historical Market Replay
Replay past sessions bar by bar with full playback controls — to study setups and practice decisions without hindsight.
Coming Soon — Targeting Late 2026
Interested in Backtest Factory?
Register your interest in the ORVINUS Trading Terminal and tell us that Backtest Factory is what caught your attention.
Registration records your interest. It does not guarantee access to testing or the product.