Inside a ₹3 L P&L: stress-testing a micro-cap equities strategy
5/17/2025
Scope: This is about risk anatomy, not secret sauce. Entry logic stays private; the focus is on how I audit any strategy before scaling.
1 · Top-line scorecard
| Metric | Value |
|---|---|
| Total realised P&L | ₹ 3.4 L |
| Daily Sharpe Ratio | 2.3 |
| Daily Sortino | 3.9 |
| Max draw-down | ₹ -1.25 L (-48 %) |
| Win-rate | 51 % |
| Average win / loss | ₹ 64 / ₹ 58 |
| Kelly fraction | ≈ 7 % of bankroll per position |
| Mean / median holding | 32 d / 8 d |
Interpretation:
- Sharpe > 2 while draw-down < 50 % keeps the sleep-at-night factor acceptable.
- Pay-off skew matters more than win-rate—coin-flip hit-rate is fine when losses are capped.
2 · Equity & pain
Left: cumulative realised P&L.
Right: draw-down stays < ₹1.3 L; time-to-recover < 25 calendar days.
3 · Trade micro-structure
Holding-time distribution
73 % of lots close inside five trading days; the tail is inventory that drifts until spreads mean-revert.
Daily P&L dispersion
- Majority of days ± ₹5 k.
- 90-th percentile winners offset the occasional -₹40 k tail.
Day-of-week edge
| Weekday | Trades | Net P&L | Avg ₹/trade |
|---|---|---|---|
| Monday | 5 % | -ve | -₹ — |
| Tuesday | 21 % | +ve | +₹ — |
| Wednesday | 22 % | flat | +₹ — |
| Thursday | 28 % | +ve | +₹ — |
| Friday | 24 % | flat | +₹ — |
(Exact rupees redacted—pattern is what matters.)
4 · Position-sizing sanity check
The simplified Kelly calc ⇒ 7 % of capital per independent bet.
I cap at 2-3 % to stay below volatility-induced margin calls; Kelly is an upper bound, not a dare.
5 · Next optimisation avenues
| Idea | Rationale |
|---|---|
| Skip Mondays | Historical bleed—Sharpe improves in sim. |
| Time-stop losers at T+2 | 50 % of negative trades rot after day 2. |
| Fee audit | Brokerage / STT shave Sharpe by ≈ 0.2—small but measurable. |
| Dynamic size on draw-down | Half-Kelly when equity < peak-10 % smooths curve without killing CAGR. |
Wrap-up
Edge without risk discipline is roulette.
Run your strategy through the same gauntlet—if your equity curve can’t live through a mark-to-market -40 %, size down or iterate until it can.