SYSTEM RUNNING · 2026
Algorithmic Trading System · NSE Equities · Intraday

1,929% return.
Zero negative years.

A rigorous historical simulation of a proprietary intraday trend-following system on NSE equities. 4+ years. Every trade accounted for. Fees deducted. No illusions.

Net PnL
₹19.3L
on ₹1L base capital
CAGR
108.1%
annualised return
Max Drawdown
12.23%
recovered in 4 days
Sharpe Ratio
5.14
world-class (>3.0)

How ₹1L became ₹20.3L

Cumulative equity across all 1,785 trades. Fixed capital mode — no compounding of position size.

2022 2023 2024 2025 2026 ₹20L ₹15L ₹10L ₹5L

Year by year

Five consecutive positive years. No calendar year ended in loss.

2022 · 241 trades
169.9%
₹1,69,870
Win rate 47.3% Fees 21.5%
2023 · 308 trades
273.6%
₹2,73,624
Win rate 51.0% Fees 17.8%
2024 · 475 trades
534.0%
₹5,33,989
Win rate 51.2% Fees 14.6%
2025 · 549 trades
722.3%
₹7,22,298
Win rate 58.3% Fees 12.7%
2026 · 212 trades (Jan–May)
230.1%
₹2,30,102
Win rate 57.5% Fees 15.0%

50 of 52 months positive

96.2% monthly win rate. Both negative months occurred in the earliest phase of the study period.

Negative
High positive

The numbers that matter

Raw returns are the story. Risk-adjusted ratios are the integrity check.

Sharpe Ratio
5.14
Benchmark: >3.0 = world-class
▲ WORLD-CLASS
Sortino Ratio
12.23
Benchmark: >5.0 = world-class
▲ WORLD-CLASS
Calmar Ratio
8.84
Benchmark: >5.0 = world-class
▲ WORLD-CLASS
Max Drawdown
12.23%
Peak to trough: 8 days
▲ RECOVERED IN 4 DAYS
₹ per ₹1 Max DD
117×
Return earned per ₹1 of drawdown
▲ EXCEPTIONAL
Win Rate
53.6%
Across 1,785 trades
▲ EDGE IS CONSISTENT

The worst it got

Maximum drawdown across the full 4+ year period. A system's resilience is measured by how it behaves under pressure — not just in favourable conditions.

Worst single period
−12.23%
₹16,456 · March 2022
Max consecutive losing days
6
Total loss: ₹22,946 (Jul–Aug 2025)
Top 3 Drawdown Episodes
#1
02 Mar – 10 Mar 2022
−12.23%
4d recovery
#2
18 Aug – 22 Sep 2022
−7.99%
77d recovery
#3
11 Sep – 15 Sep 2023
−6.17%
34d recovery
Granularity Total Periods Negative Negative % Reading
Years50 0.0% No year ended in loss
Months522 3.8% Both in early 2022
Weeks22659 26.1% Normal for intraday
Days~588~212 ~36.1% Edge compounds at month/year

What it does and why

High-level thinking. The formula is proprietary. The architecture is here.

Core Mechanism
The system is built around a proprietary trend-following filter that adjusts its sensitivity to market conditions dynamically. Unlike fixed-parameter approaches, it recalibrates how aggressively it tracks price based on recent volatility — staying in trending moves longer while tightening during choppy, low-conviction regimes.
Exit Logic
A layered exit structure sits on top of the trend signal. It is not a fixed percentage stop — it is a function of recent price behaviour. The result is that losing trades are closed faster in high-volatility environments, limiting the worst daily losses to under ₹15K even when markets moved aggressively.
Portfolio Construction
Capital is deployed across a concentrated intraday portfolio on NSE. Positions are selected for their trending behaviour and liquidity profile. Allocation across the portfolio is structured to limit single-position concentration without diluting the edge through over-diversification.
Type Intraday trend-following
Exchange NSE · Cash equities
Portfolio Concentrated intraday · 5× leverage
Avg trades/day ~3 per session
Fees paid (4+ yrs)
₹3,42,355 15.1% of gross · all figures on this page are net
Min capital required ₹1,03,656 (exact historical floor)
System Class

This system belongs to the trend-following family — a well-documented class of systematic strategies that profits from persistent directional moves in price rather than mean reversion. The edge does not come from predicting where price will go. It comes from a simple structural asymmetry: let winning trades run, cut losing trades quickly. Over enough occurrences, a win rate above 50% combined with winners larger than losers compounds into the return profile shown above. The system makes no forecast. It reacts, sizes, and exits — systematically, every session.

Beyond the algorithm

The signal is one part. What sits around it is the other.

01
Signal Layer
A proprietary algorithm monitors market conditions continuously during trading hours. When criteria are met, a signal is generated — long or short — with all trade parameters computed at that instant. No discretion, no delay, no override.
02
Multi-Broker Execution
The signal feeds simultaneously into a live execution engine connected to multiple broker accounts. Orders are placed programmatically across all accounts at signal time. The system is broker-agnostic — accounts can be added, removed, or reconfigured without touching the signal logic.
03
Per-Account Configuration
Each account operates under its own independently defined parameters — capital allocation, position sizing, instrument mapping, and signal routing. These are computed dynamically at signal time, not hardcoded. One signal, multiple accounts, each executing correctly for its own context.
04
Full Automation
From signal detection to order placement to position exit — the system requires no manual intervention during market hours. It runs, monitors, acts, and logs entirely on its own. Human involvement is limited to system maintenance and configuration changes outside market hours.
The performance shown on this page reflects the algorithm only. The infrastructure around it — multi-account execution, dynamic configuration, full automation — is what makes it operational at scale.

What you should know

Context for sophisticated readers. Not disclaimers — just the full picture.

The full picture
  • The performance shown is drawn from a rigorous backtest — a high-fidelity historical simulation across 4+ years and 1,785 trades. Every fee, every losing trade, every drawdown is in the record. Nothing is cherry-picked. A live track record is actively being built on top of this foundation.
  • Fees are fully accounted for across every trade — ₹3,42,355 paid to brokers and exchanges over the study period. All figures on this page are net. Slippage and market impact are not modelled, as is standard practice in backtesting intraday strategies.
  • This is a trend-following system. It is designed to perform in directional markets and does so with consistency — 0 negative years, 96.2% positive months. Like all trend-following strategies, mean-reverting or range-bound regimes present a different environment. The system's adaptive design is built with exactly this in mind.
  • Leverage is employed and sized deliberately. The historical capital floor of ₹1,03,656 — the exact minimum needed to have survived every losing stretch in the 4+ year record — reflects how the system is designed to be run: with adequate capital, not at the edge of margin.
  • This is not SEBI registered investment advice. This page exists to document a methodology, demonstrate technical rigour, and present a transparent performance record — nothing more.