0DTE SPXW Iron Condor Research Snapshot

What this page is meant to prove

This is not a performance pitch. It is a public example of how the desk promotes or rejects a strategy family.

The family studied here is a 10:30 New York time SPXW 0DTE iron condor workflow. The useful takeaway is not "short premium works." The useful takeaway is that the raw version, the more adaptive version, and the fixed-rule versions do not behave the same way.

Research design

• Instrument family: SPXW 0DTE iron condors
• Entry snapshot: 10:30 New York time
• Sample window: 2024-01-02 through 2026-03-24
• Input trade days: 551
• Walk-forward design: 7 out-of-sample folds
• Promotion rule: only fixed, interpretable rules are candidates for live use

The broader proof standard behind the research repo is simple: identical data hygiene, identical cost assumptions where possible, explicit failure reasons, full-sample results first, then fixed-rule walk-forward evaluation.

What the raw family did

The raw 10:30 standard condor variant produced these full-sample results on 551 priced trades:

Metric	Value
Total P&L per 1-lot condor	-$480.04
Average P&L per trade	-$0.87
Positive P&L rate	88.57%
Implied breakeven win rate	89.67%
Max drawdown	-$5,129.02

That is exactly why win rate alone is not enough. A strategy can win often and still fail the economics of the payout profile.

What survived out of sample

The walk-forward layer is more important than the raw full-sample read.

Variant	Sample	Trades	Total P&L	Comment
Baseline walk-forward	7 test folds	425	+$1,998.97	4 of 7 test folds positive
Dynamic selector	7 selected folds	164	-$1,160.03	Rejected
Fixed rule: downside-adaptive farther-put condor	6 folds	178	+$1,511.00	83.3% positive test folds
Fixed rule: skip high put-share days	6 folds	137	+$1,649.99	Worst test fold -$197.02

Two details matter here:

1. The dynamic selector did not survive out of sample, even though it was more adaptive and more complex.
2. The fixed-rule variants that remained interpretable held up better than the "smart" selector.

Plain-English version of the promoted rules

1. Downside-adaptive farther-put variant

When the band is wide enough, the available credit clears a minimum floor, and downside stress is visible through a risk-off regime plus elevated put-side credit share, the workflow shifts from the standard structure to a farther-put condor.

Public takeaway: on jumpy downside days, the response is not automatically "sell the same condor anyway." The structure itself changes.

2. Skip high put-share days

When downside pressure is too concentrated in the put side, the better decision can be to skip the setup entirely.

Public takeaway: sometimes the edge is not in finding a cleverer trade. It is in refusing the wrong one.

Why this is public

This is the kind of research proof the site should carry:

• a concrete instrument family
• explicit sample period
• clear full-sample and walk-forward reads
• a rejected adaptive layer
• interpretable rules that remain after promotion

That is closer to how the desk actually works than a generic "strategy works" claim.

What this does and does not say

It does say that the desk treats volatility-selling research as a measured research program with promotion and rejection rules.

It does not say that the public should trade this structure, that the historical result will persist, or that short-dated index options are suitable for any specific person or mandate.