Share Buybacks— repurchase-announcement tilt

Companies that announce a share-repurchase program (8-K filed with the SEC) tend to modestly out-drift the market over the following year. Each trading day we scan EDGAR for new repurchase 8-Ks across the S&P 500+400, enter the name at the next close (no lookahead), hold ~252 trading days, and mark the equal-weight basket live against SPY. A long-only tracking-error tilt, forward-tracked in shadow mode (no orders).

Backtest (2015-2023, EDGAR 8-K, monthly basket)

Excess vs SPY
+4.8%/yr
Underwrite
+3.0%/yr
Info ratio
0.71
Hit rate
63%
Excess maxDD
-11%
Corr to SPY
0.94

⚠️ Long-only tilt, corr 0.94 to SPY — the edge is the +3-5%/yr tracking-error (excess-over-SPY) stream, NOT market-neutral. Backtest carries ~1-2pt survivorship bias; underwrite ~+3%/yr excess. Paper/shadow only — no orders.

Live paper book (current basket vs SPY)

Basket excess
+0.10%
Since inception
-0.70%
Open positions
445
Avg open excess
-4.31%
Added since incept.
103
Days live
65

Excess-over-SPY equity of the equal-weight basket. Live paper book; excess = basket total-return minus SPY. Reconstructed pre-inception positions use each name's real post-filing entry price (no lookahead). Data asof 2026-09-18.

Top current holdings (excess since entry)

TickerEntryDaysExcess vs SPY
MU2025-09-24359+513.0%
MRVL2025-09-25358+175.3%
CRWD2026-03-04198+121.4%
MU2026-03-19183+112.7%
TWLO2026-02-13217+103.5%
PANW2026-01-30231+94.8%
OKTA2026-01-06255+91.1%
HAE2026-02-06224+71.8%
ILMN2026-05-01140+70.5%
NTNX2026-04-08163+64.2%
DINO2026-05-20121+61.0%
ROKU2026-02-13217+58.3%
KEYS2025-11-25297+57.6%
HALO2026-05-12129+55.1%
FDX2025-09-19364+49.6%

Equal-weight basket of S&P 500+400 names filing an 8-K share-repurchase program in the trailing 12 months; each entered at its post-filing close (lookahead-free), held 252 trading days, marked daily vs SPY.