The rules
Rodier, Haryanto, Shum and Hejazi (October 2012 working paper, as summarized by CXO Advisory; published in the Review of Finance in 2016). If the market is up at least 2% as of 14:45 ET, buy SSO (ProShares Ultra S&P500, a leveraged fund that aims for twice the S&P 500's daily return); if it is down at least 2%, buy SDS (ProShares UltraShort S&P500, which aims for twice the opposite of that return); otherwise do not trade. Sell at the close. A second version moves the signal to 14:30 and adds an estimate of the leveraged funds' rebalancing.
- Universe
- SPY
- Bars
- 1-minute
- Position size
- 97% of equity in SPY, bought or sold short. Published: the full account in SSO or SDS, both leveraged 2x funds.
- Rebalance
- Only on full sessions when SPY is 2% or more above or below its prior close at 14:45, held to the close. Published: days the market was up or down 2% or more as of 14:45, held to the close; 128 trades over 2006-2011.
What the research found
Cheng and Madhavan (Journal of Investment Management 2009) show leveraged funds resetting daily must trade with the day's move near the close. Shum, Hejazi, Haryanto and Rodier (Review of Finance 2016; 346 large stocks, 2006-2011) find end-of-day volatility rising with potential rebalancing, with impacts not all large enough to matter (economically significant), but largest on the most volatile days. Their price-only rule, in CXO Advisory's summary, made 128 trades averaging 0.60% before costs on the 2x funds. Profits came mostly from 2.5 months of late 2008; CXO puts break-even costs at about 0.6% a round trip and warns that the rules were picked on the same data they were tested on, so results likely overstate what to expect. Lenkey's 2024 survey judges the published paper's 14:30 version (0.66% a trade before costs) not significantly different from its 2% rule (0.60%), and says that version picks its days with information not known at 14:30. On equity index futures (contracts that track a stock index), Baltussen, Da, Lammers and Martens (Journal of Financial Economics 2021) find last-half-hour momentum stronger where leveraged funds are a larger share of the market (2006-2020), and continuation to the close once the return since the prior close, checked at each half hour, passes the 10th or 90th percentile of daily returns (1982-2020). Against: Ivanov and Lenkey (Journal of Financial Markets 2018; 2006-2014) find investor flows into and out of the funds offset much of the rebalancing, leaving an effect on late-day returns too small to matter even on the busiest rebalancing days, and Lenkey's survey, by a coauthor of that paper, agrees. Barbon and coauthors (working paper 2022) find leveraged-fund rebalancing still pushing stocks in its direction in the last half-hour over 2012-2019, reversed by the next open, but declining to barely significant late in the sample. QuantRocket's 2019 test found a related rule (long or short a leveraged fund after its own move of more than 2% from the prior close at 14:00, held to the close; about 0.67% on the index for its mostly 3x funds) profitable on 14 funds until mid-2016, then declining; with a 6% threshold, about 2% on the index (derived), it grew 31% a year after costs over 2008-2016 before flattening from 2017. The published edge is 0.60% a trade before costs on the 2x funds; halved for 1x, about 0.30% (derived), against the 0.10% the library charges a round trip.
Results
| Measure | Orbital test | SPY, same days |
|---|---|---|
| Window | Oct 30, 2023 to Oct 30, 2026 | Oct 30, 2023 to Oct 30, 2026 |
| Total return | −0.90% | +88.60% |
| CAGR | −0.30% | +23.54% |
| Sharpe | −2.76 | 1.22 |
| Discounted Sharpe | −6.79 | — |
| Max drawdown | −3.6% | −18.7% |
| Win rate | 47.8% | — |
| Profit factor | 1.33 | — |
| Trades | 23 | — |
Hypothetical: a backtest with modeled costs, not traded with real money.
Hypothetical growth of $10,000
R12 Big-Move Late-Day Momentum · −0.9% vs SPY +88.6%
Rebuild it in Orbital
Paste this into Orbital Chat. The Strategy Developer builds the workflow and asks before it runs any backtest.
Build a program on 1-minute bars for SPY, using the session clock. Keep the prior session's split-adjusted close: the open of its 16:00 closing-print minute, or on early-close days the close of its 13:00 minute, or its last close if that minute has no bar. When the 14:44 bar closes, if that bar's close is 2% or more above the prior close, buy SPY with 97% of equity; if it is 2% or more below, sell SPY short with 97% of equity; otherwise do nothing that day. Either order fills at the 14:45 open. Close the position when the 15:59 bar closes, so it fills at the open of the 16:00 closing-print minute. Take no trade on the backtest's first session, which has no prior close, or on early-close days. Use a margin account so it can sell short. Backtest the last 3 years with $100,000.
Is it real?
- Strategies tested together
- 10
- Beta to SPY
- 0.03
- Alpha vs SPY
- −5.0% a year
- Costs paid
- $2,200, 169% of the gain before costs
- Data
- 1m decisions on 1-minute bars, 754 sessions
- Fills
- Next open at market, sized at the fill
Read this before you trust it
SPY at 1x replaces SSO and SDS, so each trade carries half the published exposure, and a down day is traded by selling SPY short on margin instead of buying the inverse fund. The sources say only that the market was up or down 2% or more as of 14:45; the build measures SPY's own move from its prior close (SPY tracks the S&P 500). After an early-close day that prior close is the 13:00 minute's close, since that day's closing auction is usually that minute's last trade. The entry fills at the 14:45 minute's open and the exit at the open of the 16:00 closing-print minute, measured by Orbital at a median 0% from the official close, rather than in the closing auction, which some of the funds use for their rebalancing. Early-close days are skipped, since 14:45 is after their 13:00 close. The 14:30 version needs the funds' assets, which Orbital does not have. 97% of equity. Each of a trade's two fills pays 0.05% slippage, an assumed trading cost.
Sources