Industries trend. Information that affects a whole sector is absorbed slowly, and institutional money rotates between sectors over weeks to months, so the sectors that have led over the past year tend to keep leading for a while (Moskowitz and Grinblatt). A long-only rotation among the 11 SPDR sector ETFs, holding only sectors above their own 200-day average, should ride that drift and step aside in bear markets.
The likely failure was written into the pre-registration: a long-only sector strategy with beta near 1, tested in a bull market, may be nothing but beta plus a sector tilt.
How it was tested
Pre-registered on 2026-06-01. The ETF history had to be downloaded separately, because ETFs are not in the lab's factor library. Signal: positive 12-1 month return, price above the 200-day average, and a 60-day momentum rank of at least 80 among the 11 ETFs. Buy at the next open, hold 21 sessions (a monthly rebalance), 3×ATR stop, no target. Alpha was measured against SPY with beta assumed equal to 1, the instrument of the time, with a bar of t ≥ 1.65.
Where it died
One run. Alpha against SPY was −0.139% per trade over 558 trades, t −0.98. Gate 8, under the capture rule of the time, gave t −1.04 on the captured trades and a capture ratio of 0. Both measurements predate the Jensen-alpha gate and the 2026-09-19 simulator slot-leak fix, and the alpha was not re-derived with Jensen alpha. The confound named in advance is the one that showed up.
The record
Thesis
Sectors with the strongest 12-1 month momentum keep outperforming laggards for the following weeks, because industry-wide information is absorbed slowly and institutional flows rotate over weeks to months. A long-only rotation among the 11 SPDR sector ETFs, only holding sectors that are themselves above their 200-day average to avoid bear-market traps.
Prior: Written into the pre-registration: a long-only, beta-about-1 sector strategy in a bull market may be nothing but beta plus a sector tilt, and if so it should fail the alpha gate.
Moskowitz, T. & Grinblatt, M. (1999). Do Industries Explain Momentum? Journal of Finance 54(4).
Faber, M. (2007). A Quantitative Approach to Tactical Asset Allocation. Journal of Wealth Management.
Specification
signal
12-1 month return > 0; above the 200-day MA; 60-day momentum rank ≥ 80 among the 11 ETFs
universe
11 SPDR sector ETFs
entry
next open
exit
21 sessions (monthly rebalance), 3×ATR stop, no target
data
ETF history downloaded separately; ETFs are not in the lab's factor library
Gates
Pre-registrationpass
Was the hypothesis, its parameters and its pass/fail rule written down before any data was looked at?
Market-adjusted alpha (legacy)fail
Does the strategy beat SPY over the same holding windows, assuming beta = 1?
alpha t ≥ 1.65
alpha t-0.98alpha / trade %-0.14trades558
Alpha vs SPY with beta assumed = 1 (the instrument of the time); not re-derived with Jensen alpha.
Gate 8 — Capture-preserved alphafail
Does the subset of trades a real portfolio can actually take keep the alpha?
t_cap ≥ 1.65 and ratio ≥ 0.30 (rule of the time)
t captured-1.04capture ratio0
Measured with the pre-2026-09 capture ratio and before the 2026-09-19 simulator slot-leak fix.
Why it died
Engine Sharpe +1.18 and +0.33% gross per trade — and −0.139% per trade against SPY over the same windows. Buying the strongest sectors in a bull market makes money in absolute terms but does slightly worse than holding SPY: beta plus a sector tilt, exactly the confound the pre-registration warned about.
What it taught
Under the lab's pre-May methodology a Sharpe of 1.18 would have looked like an edge. Writing the likely failure mode into the pre-registration made the verdict a confirmation instead of a disappointment, and closed long-only sector momentum on SPDRs as a retail alpha source.
Verdict history
2026-06-01Pre-registered; ETF history downloaded after an infrastructure blocker
2026-06-01Single run: alpha vs SPY t −0.98, Gate 8 t_captured −1.04 Beta in disguise