Vertex Algo — Options exit strategy backtester

The problem

Most algos lose money on the stop, not the signal.

Traders spend months tuning entry ideas and market timing, then slap on a stop-loss not fitted to the stratagys niche movements.

TOO TIGHT

Whipsawed out

A stop sized for calm markets gets clipped by normal noise, turning winning trades into losses before the move even starts.

TOO LOOSE

Death by drawdown

A stop sized for volatile regimes bleeds capital in quiet ones, dragging down Sharpe and eroding trust in the system.

STATIC

Blind to regime shifts

One stop distance rarely fits pre-earnings chop, trending momentum, and low-liquidity overnight sessions equally well.

What the engine tests

80+ exit strategies, replayed bar-by-bar on live options data.

Paste an OCC-format options contract with an entry time and the engine fetches 1-minute bars, then walks every bar applying each strategy exactly as it would have fired live — no lookahead, no approximations.

Strategy family

How it's calculated

Variants tested

What to watch for

Bracket (TP + SL)

Take profit at +X%, stop loss at −Y% from entry

TP50/SL15, TP100/SL25, TP25/SL15 and more (12 R:R combos)

R:R ratio vs actual win rate on your contracts

Trailing stop

Highest close since entry − trailing offset (% or $)

10%, 15%, 20%, 25%, 30%, dollar-based

Gives back open profit — too tight = whipsawed

ATR / volatility stop

Entry − (N × ATR) or stddev envelope, recalculated per bar

ATR trail, Bollinger stop, DevStop, vol spike exit

Lookback window sensitivity; wide in high-vol regimes

Greek-based exit

Exit when delta, vega, theta, charm, or vanna cross a threshold — computed from live Black-Scholes per bar

Delta 0.30 / 0.40 / 0.50, delta trail, vega limit, charm, vanna, theta rate

Options-native — reacts to the actual risk profile, not just price

IV / VIX trigger

Exit when implied volatility or a VIX proxy crosses a regime threshold

VIX trigger, vol skew change, calendar IV diff, IV crush

Volatility mean-reversion after events; avoids holding through crush

VWAP / underlying move

Exit on VWAP cross, underlying % move, underlying ATR, or technical level breach

VWAP stop, underlying pct move, underlying dollar move, underlying ATR

Tracks the underlying's price action rather than the option premium

All 80+ strategies — head to head
Same contracts, same fills, strategy logic swapped per run
Every strategy across every contract in one run
Full leaderboard: Sharpe, total P&L, win rate, profit factor, exit reason breakdown

How it works

From strategy to stop-loss report in three steps.

Connect

Paste your options contracts

Enter OCC-format contracts with entry time — one per line (e.g. SPY260112C00690000, 09:30). The engine fetches 1-minute bar data automatically. No CSV, no manual data prep.

Simulate

Every exit strategy runs simultaneously

80+ strategies — bracket TP/SL, trailing %, ATR, VWAP, delta thresholds, IV triggers, greek-based exits — replay bar-by-bar against the same bars and fills. Greeks are computed live per bar where needed.

Decide

Leaderboard + AI analysis

Full ranking table sorted by Sharpe, total P&L, win rate, or profit factor. Exit reason breakdown per strategy. Overlay any stop trace on the price chart. Hit the AI Summary button for an instant narrative analysis of what worked and why.

Try it yourself

Same entry. Same bars. Five different exits.

Three real SPY contracts from April 2026 — same fills, same data, only the exit rule changes. Pick a scenario to see how stop selection alone moves the outcome.

▲ WIN

SPY 690C

+$80

Tariff pause rally — the trailing stop got shaken out early, missing the $80 move

▼ LOSE

SPY 680P

−$98

SPY kept ripping — fast exits saved $200 vs waiting for a reversal

─ CHOP

SPY 660P

+$55

Post-bounce consolidation — structured take-profit outperformed trailing 2.5×

Option price & stop levels

SPY underlying

5 stop strategies — same entry, same bars, different rules

Learn

A backtester teaches you more when you can see the math behind it.

Every subscription includes access to the Vertex Algo Discord — where the stop-loss math and the code that implements it are broken down together, not sold as a black box.

Code & math channels

Ongoing, not one-and-done

Dedicated channels for posting your own backtest code, walking through a strategy's math, and getting feedback from other members working on similar problems.

Live sessions

Office hours on the engine itself

Regular live walkthroughs of the backtester's internals — reading a stop-loss report line by line, or debugging why a strategy's numbers don't match expectations.

Included with your subscription

Discord access is bundled into every Desk subscription.

# chandelier stop, worked in #code-and-math def chandelier_stop(high, atr, mult=3): return high.cummax() - (mult * atr)

Your entries are fine.
Your exits are costing you.