Help us double down on what's working, instead of guessing. Takes 5 seconds, totally optional.
research · 6 min read
Does DEX Predict QQQ? A 9,216-Strategy Options Backtest
We backtested 9,216 QQQ option strategies on dealer delta exposure (DEX) and the opening gap, 2022 to 2026. None beat random timing; the best Sharpe was 1.01.
Short answer: no. Buying QQQ calls after a gap up with positive DEX lost money on average in all 24 delta and expiry setups we tested, from 3 January 2022 to 19 October 2026. We then tested 9,216 variations: 24 DEX and gap rules, calls, puts and debit spreads, four exits, six deltas and four expiries. The best Sharpe ratio was 1.01. Shifting the same rules to random dates typically produces a best of 1.11, so no variation beat chance.
24 of 24setups of the original trade lost money on average
1.01best Sharpe ratio of 9,216 strategies; random timing typically reaches 1.11
0 of 8,144strategies with 60+ trades that beat random timing
34 vs 142+positive-DEX sessions in 2022 vs each year since: DEX sign tracks the market
What we tested
The idea is simple. If dealer delta exposure (DEX) was positive at yesterday's close and QQQ opens higher, buy a call at 09:35 ET and sell it before the close. We tested that trade at six deltas (0.10 to 0.50) and four expiries (1 week to 3 months), then widened the search:
24 entry rules: gap up or down, positive or negative DEX, stronger DEX readings, larger gaps, DEX that rose or flipped sign, DEX confirmed at 09:34, DEX ranked against the previous 60 sessions, and price momentum.
4 payoffs: long call, long put, bull call spread and bear put spread.
4 exits: same day, next day, after 5 sessions, or at expiry.
That is 24 × 4 × 4 × 6 × 4 = 9,216 strategies over 1,203 sessions. Every strategy opens one lot per signal and pays the ask to buy and the bid to sell, plus fees.
The original trade at every delta and expiry
Each cell is the Sharpe ratio of one setup: buy a QQQ call at 09:35 ET after a gap up with positive DEX, sell at 15:55 ET. Every setup lost money on average, from −10.05% to −0.07% per trade. Use the menus to see any other rule, payoff or exit; select a cell to chart it below.
Gap up + positive DEX, long call, same-day exit: Sharpe ratio (average return per trade and trade count below)
Delta
1 week
1 month
2 months
3 months
0.10
−1.07avg −10.1% · 362 trades
−0.85avg −3.5% · 422 trades
−0.61avg −2.1% · 329 trades
−1.21avg −3.2% · 345 trades
0.20
−0.43avg −3.0% · 363 trades
−0.42avg −1.4% · 429 trades
−0.27avg −0.5% · 331 trades
−0.52avg −1.2% · 352 trades
0.25
−0.47avg −2.5% · 367 trades
−0.36avg −1.1% · 426 trades
−0.20avg −0.2% · 328 trades
−0.42avg −0.7% · 346 trades
0.30
−0.34avg −1.5% · 380 trades
−0.32avg −0.7% · 427 trades
−0.20avg −0.1% · 332 trades
−0.39avg −0.6% · 350 trades
0.40
−0.17avg −0.8% · 373 trades
−0.21avg −0.6% · 432 trades
−0.15avg −0.1% · 332 trades
−0.26avg −0.4% · 350 trades
0.50
−0.21avg −1.0% · 369 trades
−0.31avg −0.7% · 431 trades
−0.20avg −0.1% · 332 trades
−0.59avg −0.7% · 349 trades
Free account
The rest of this study is free with an account
A free account opens all 9,216 strategies ranked, the profit chart for any strategy, the luck test and the year-by-year breakdown.
Data: FlashAlpha historical QQQ option quotes, stock quotes and DEX, 1,203 sessions from 3 January 2022 to 19 October 2026.
DEX: net DEX divided by total absolute DEX across all strikes and expiries, from /v1/exposure/dex/QQQ at 15:55 ET on the previous session. Positive means dealers were net long delta.
Gap: the 09:31 ET QQQ quote midpoint against the previous 16:00 ET midpoint.
Entry: 09:35 ET. Expiry nearest the target within 4 to 10, 25 to 35, 53 to 67 or 80 to 100 days, then the out-of-the-money option nearest the target delta, within 0.05. Open interest at least 250, bid/ask spread at most 20%, quote no more than five minutes old.
Spreads: sell a further out-of-the-money option at half the target delta in the same expiry.
Fills and costs: buy at the ask + $0.01, sell at the bid − $0.01, $1 per contract per trade. Also tested at $0.05 and $0.10 slippage.
Exit: 15:55 ET (12:55 on early-close days) on the same day, the next session, the fifth session, or expiry day. Trades still open on 19 October 2026 are left out.
Sharpe ratio: monthly realized P&L, one lot per signal, annualized, no risk-free rate. Open positions are not marked to market.
Not modeled: early assignment, taxes, market impact and API costs.
How much to trust it
It covers one period of under five years and one symbol.
Strategies share dates and contracts, so 9,216 results are not 9,216 independent tests.
The luck check was designed after the results were seen. It is not an out-of-sample test.
Run your own DEX backtest
Every number here comes from FlashAlpha endpoints read at the moment of each decision. On the Alpha plan, the Historical API replays historical DEX, option quotes and the rest of the exposure suite at any past minute, so you can test your own rules, deltas and exits. For live readings, see the DEX endpoint and plans.
Cite this research
FlashAlpha Research. "Does DEX Predict QQQ? A 9,216-Strategy Options Backtest." 22 October 2026. https://flashalpha.com/articles/does-dex-predict-qqq-delta-exposure-options-backtest
Frequently asked questions
Does positive DEX predict QQQ going up?
Not in a way you can trade. Buying QQQ calls after a gap up with positive DEX lost money on average in all 24 delta and expiry setups from 3 January 2022 to 19 October 2026. The 0.25-delta, 1-month version averaged −1.1% per trade over 426 trades, with a Sharpe ratio of −0.36.
What is DEX (delta exposure)?
DEX is the net delta that options dealers hold across all strikes and expiries. FlashAlpha reports it as net_dex; positive DEX means dealers are net long delta. This study divides net DEX by the total absolute DEX so that days can be compared.
Which DEX strategy had the best Sharpe ratio?
Of 8,144 strategies with at least 60 trades, the best was Gap up + DEX in upper prior-60-session quartile: long call, 0.50 delta, 1 week, same-day exit, with a Sharpe ratio of 1.01. Random timing typically produces a best Sharpe of 1.11 across the same search, so this result is no better than chance (adjusted p = 0.69).
Why do negative-DEX strategies look better than positive-DEX ones?
Mostly because of when they happened. 217 of the 414 negative-DEX sessions were in 2022, a falling market. Positive DEX occurred on only 34 sessions in 2022 and on 142 to 224 sessions a year in the 2023 to 2026 rally. Comparing the two signs mostly compares years.
How did you check whether a result was luck?
Each rule's dates were shifted along the calendar 1,078 times. A shifted rule fires just as often and in the same clusters, but on the wrong days. The best Sharpe among all real strategies was compared with the best Sharpe each shifted run produced. 0 of 8,144 strategies beat that test.
Can I run this DEX backtest myself?
Yes. The DEX values and option quotes come from FlashAlpha's Historical API, which replays endpoints at any past minute in the archive. It is available on the Alpha plan.
Positive DEX followed by a gap up did not make QQQ calls profitable from 2022 to 2026, at any delta or expiry. Across 9,216 variations, the best Sharpe ratio (1.01) was lower than random timing typically produces, and DEX sign mostly tracked whether the market was in the 2022 decline or the rally after it.
