DeployQuantSGOV trading strategies, backtested
iShares 0-3 Month Treasury Bond ETF: T-bills, often used as the cash leg of rotation strategies. Every DeployQuant template run on SGOV over 5.7 years of minute data, same engine, same window, sorted by return.
SGOV holds Treasury bills with up to three months to maturity. It is the closest thing in this test to cash that pays interest, and the data shows it: annualized volatility of 0.24%, a deepest drawdown of 0.02%, and a price series that rose 20.47% in total from 2021-01-04 to 2026-10-02. Buy-and-hold made 3.22% a year and turned $10,000 into $11,996.
Twelve templates ran on SGOV. None beat buy-and-hold on CAGR. Eight finished with a positive CAGR and four had a shallower drawdown than the fund's 0.02%. The first six in the ranking sit between 3.20% and 3.22%, which is the fund's own return reached through different amounts of time invested. The four at the bottom made 0.00% or 0.02% because they were out of the fund, and out of the fund means in cash earning nothing in this test.
This page is a control case. A trading rule needs price movement to work with, and SGOV supplies almost none. What the results show is what a rule costs when there is nothing to gain: the fund's return is a slow, steady accrual, and every extra trade can only take away from it.
The headline run has no fees or slippage. The cost runs at 5 and 10 basis points matter more here than on any volatile fund, because a basis point is a large share of a day's return. Those are covered in the trading section.
| Strategy | CAGR | max DD | Sharpe | trades | win rate | final value |
|---|---|---|---|---|---|---|
| Golden Cross (SMA 50/200) | 3.2% | −0.0% | 14.13 | 0 | – (+1 open) | $11,993 |
| 200-Day SMA Regime Filter | 3.2% | −0.0% | 14.13 | 5 | 40% (+1 open) | $11,997 |
| SMA-200 Trend + 15% Trailing Stop | 3.2% | −0.0% | 14.13 | 0 | – (+1 open) | $11,993 |
| SMA 10/50 Trend | 3.2% | −0.1% | 13.91 | 6 | 33% (+1 open) | $11,990 |
| EMA 12/26 Trend | 3.2% | −0.1% | 13.86 | 17 | 47% (+1 open) | $11,991 |
| Weekly Entry + 7% Target | 3.2% | −0.0% | 13.79 | 3 | 67% (+1 open) | $11,984 |
| First-to-Last Day of Month | 3.0% | −0.1% | 13.00 | 69 | 90% (+1 open) | $11,836 |
| RSI(2) Dip Snapback | 0.0% | −0.1% | 0.42 | 52 | 46% | $10,012 |
| RSI(14) Mean Reversion | 0.0% | −0.0% | 0.00 | 0 | – | $10,000 |
| Drawdown Dip Buyer + 8% Target | 0.0% | −0.0% | 0.00 | 0 | – | $10,000 |
| 20-Day Momentum + Trailing Stop | 0.0% | −0.0% | 0.00 | 0 | – | $10,000 |
| 3-Month Momentum Switch | 0.0% | −0.0% | 0.00 | 0 | – | $10,000 |
| Buy & hold SGOV | 3.2% | −0.0% | 13.84 | – | – | $11,996 |
A ranking decided by time in the fund
The top of the table is effectively a tie. The golden cross, the 200-day regime filter and the trend plus trailing stop each made 3.22%. The SMA 10/50 trend and the EMA 12/26 trend made 3.21%. The weekly 7% target made 3.20%. The ending values are $11,993, $11,997, $11,993, $11,990, $11,991 and $11,984, against $11,996 for holding. The differences are a few dollars on $10,000 and come from the days each rule was out of the fund.
The golden cross and the trend plus trailing stop each made one fill and held a single position from 2021-10-19 at an adjusted 83.40, still open at the end with a gain of 20.43%. Both were invested on 86.1% of days, because the 200-day average needed 200 sessions before the signal could fire. The 200-day regime filter entered the same day, then made 5 closed round trips in October 2021 and January 2022 with a profit factor of 3.00, and opened its final position on 2022-01-26 at 83.39. Its exposure was 85.7%.
The SMA 10/50 and EMA 12/26 templates traded more. SMA 10/50 made 6 round trips, two of them wins at 0.00%, with a longest losing streak of 6 and a profit factor of 0. Its worst trades lost 0.02% and 0.01%. EMA 12/26 made 17 round trips with a 47% win rate, an average win of 0.01% and a profit factor of 0.75. These trades are measured in hundredths of a percent, and their sizes are the width of a bid-ask spread more than a market move. In both cases the 2021 column holds almost all the trades, 6 for SMA 10/50 and 15 for EMA 12/26, and the later years have none or two. Once the fund's price began to climb steadily after 2022, the averages stopped crossing.
The weekly 7% target has the strangest trade log. It made 3 round trips and won 2, with an average win of 7.02%, a best trade of 7.03% and a profit factor of 1,211. The winners are the 1095-day hold from 2021-01-11 to 2024-01-11 at 7.03%, and the 512-day hold from 2024-01-16 to 2025-06-11 at 7.00%. The 7% target on a T-bill fund is reached by the slow accrual of income, over years, and the average hold was 536.7 days. A strategy designed for weekly swings was a buy-and-hold with a long wait. Its exposure was 99.5%. A third position opened on 2025-06-16 at 95.58 was still open at the end with a gain of 5.08%.
The monthly cycle made 2.98% with 69 round trips and a 90% win rate, and a profit factor of 314.27. It won 62 of 69 trips and had a winning streak of 55. The average win was 0.29% and the average loss was -0.01%. The best trade was 0.45% from 2023-08-01 to 2023-08-31, the month the fund had its best month, 0.49% in buy-and-hold. The 2021 results show 4 wins out of 12, the year when the fund returned 0.05%. Exposure was 95.2%, and it trailed the fund by the days it spent out at the start and end of each month.
The RSI(2) snapback made 0.02% from 52 round trips with a 46% win rate, 7.1% exposure and an average hold of 3.1 days. Its profit factor of 1.71 sounds good and describes a total return of 0.12%, with end equity $10,012. The fund hit an RSI(2) below 10 on 17 sessions only, but the template still found 52 entries, so its entries do not map onto that count exactly.
The last four templates made no trades. The RSI(14) mean reversion, the dip buyer, the momentum breakout and the 3-month momentum switch each finished with 0 fills, 0.00% CAGR, 0.00% drawdown and $10,000. The data bears out why two of them could not fire. RSI(14) fell below 30 on 0 sessions and the fund had 0 falls of 10% or more from a 20-day high. Zero return is the cash value of those rules, and the gap to holding is 3.22 points of CAGR a year.
How the years and drawdowns separate the templates
The calendar-year columns show where each template lost its few basis points. Buy-and-hold made 0% in 2021, 1.5% in 2022, 5% in 2023, 5.2% in 2024, 4.1% in 2025 and 2.7% in 2026 to date. The golden cross, the trend plus trailing stop, the SMA 10/50 and the EMA 12/26 matched those figures in every year, so their small CAGR gap to the fund is invisible at one decimal. The 200-day regime filter is the only template that beat the fund in any year: it made 1.6% in 2022, against 1.5% for holding. Its trades that year were the round trips in January 2022, and the gap is a tenth of a point.
The monthly cycle shows the cost of its boundary days most clearly. It trailed the fund by 0.1 points in 2022, 0.3 in 2023, 0.5 in 2024, 0.2 in 2025 and 0.2 in 2026 to date. The gap is largest in the years when the fund earned the most, which fits a rule that is out of the fund on a fixed share of days. Its worst month was October 2021 at -0.02%, and its best was August 2023 at 0.44%, against 0.49% for holding in the same month.
The RSI(2) snapback lost 0.1% in 2021 and made 0.2% in 2022, then rounded to 0% in 2023 to 2026. It was in the fund 7.1% of the time, so in the years when the fund paid about 5%, the rule collected almost none of it. Its worst month was December 2021 at -0.05% and its best was January 2022 at 0.05%. Its 0.08% drawdown ran 301 days from the 2021-03-04 peak to the 2021-12-30 trough, then took 78 days to recover. That is the longest slide of any template relative to its size, and it happened in the one year when the fund itself paid nothing.
The drawdown lists point the same way. The fund's three deepest drawdowns were each 0.02%, on 2021-03-03, 2021-08-02 and 2021-08-12, and each recovered within 4 sessions. The SMA 10/50 and EMA 12/26 drawdowns reached 0.05%, taking 77 and 82 days to hit bottom after the 2021-08-04 peak and recovered on 2022-03-30 in both cases. These are the whipsaws of October 2021 playing out in a series that barely moves. The weekly 7% target trailed by 0.1 points in 2024. Its first position was sold on 2024-01-11 at 89.23 and the next was bought on 2024-01-16 at 89.27, a slightly higher price.
How each strategy traded SGOV
| Strategy | Time in market | Avg hold (days) | Best trade | Worst trade | Profit factor | With 10 bps slippage |
|---|---|---|---|---|---|---|
| golden cross | 86.1% | – | – | – | – | 3.2% |
| 200-day regime filter | 85.7% | 18 | 0.0% | −0.0% | 3.00 | 3.0% |
| trend + trailing stop | 86.1% | – | – | – | – | 3.2% |
| SMA 10/50 trend | 93.5% | 29 | 0.0% | −0.0% | – | 3.0% |
| EMA 12/26 trend | 95.0% | 18 | 0.0% | −0.0% | 0.75 | 2.6% |
| weekly 7% target | 99.5% | 537 | 7.0% | −0.0% | 1211.00 | −4.8% |
| monthly cycle | 95.2% | 28 | 0.5% | −0.0% | 314.27 | 0.6% |
| RSI(2) snapback | 7.1% | 3 | 0.0% | −0.0% | 1.71 | −1.7% |
| RSI mean reversion | 0.0% | – | – | – | – | 0.0% |
| dip buyer | 0.0% | – | – | – | – | 0.0% |
| momentum breakout | 0.0% | – | – | – | – | 0.0% |
| 3-month momentum | 0.0% | – | – | – | – | 0.0% |
What costs do to a fund that moves a cent a day
SGOV moves so little that a 5 basis point cost is large compared with a day's return. The cost runs show how large.
The weekly 7% target made 3.20% in the headline run. At 5 basis points it made 1.08% with a max drawdown of 8.23%, and at 10 basis points it lost 4.85% with a max drawdown of 25.34%, ending at $7,515. That is a 25.34% drawdown on a fund whose largest drawdown was 0.02%. The template has 7 fills in the headline run, and the cost is charged on each. A fixed percentage cost on a fund with 0.24% annual volatility cannot be recovered.
The RSI(2) snapback went from 0.02% to -0.86% and -1.73% with 104 fills, ending at $9,045 at the higher cost. Its max drawdown rose from 0.08% to 9.55%. The monthly cycle went from 2.98% to 1.77% and then 0.58%, and its Sharpe ratio fell from 13.00 to 5.13 and then 1.10. The EMA 12/26 trend with 35 fills went from 3.21% to 2.91% and 2.60%, with its drawdown rising from 0.05% to 3.38%. The SMA 10/50 trend went from 3.21% to 3.09% and 3.00%. The 200-day filter went from 3.22% to 3.12% and 3.04%.
The golden cross and the trend plus trailing stop barely changed, from 3.22% to 3.21% and 3.20%, because they made one fill each. The templates that did nothing are unchanged at $10,000. For this fund the cost run reads as a count of fills: more fills, larger loss.
The Sharpe ratios in the table are very large and should not be read as skill. Buy-and-hold has a Sharpe of 13.84 and the golden cross 14.14, because the volatility in the denominator is 0.24% a year. The ratio measures how smooth the accrual was, and the fund was smooth. Several templates with a Sharpe above 13 had the same return as holding.
Liquidity is a separate point. The fund trades about $532,191,384 a day on average and the median minute carries 8,376 shares, so a cost of 5 or 10 basis points is probably high for this fund in practice. The test does not measure that, and it applies the same cost assumption to every ETF.
Hold times show the usual split. The golden cross and the trend plus trailing stop held a single position for the whole window after entry. The 200-day regime filter held 18.2 days on average, SMA 10/50 held 28.7, EMA 12/26 17.8, the monthly cycle 28.4, the weekly template 536.7 and the RSI(2) 3.1.
How SGOV behaved
| Measure | SGOV |
|---|---|
| Data in this test | 2021-01-04 to 2026-10-02 (1444 sessions) |
| Total return, buy and hold | 20.5% |
| Annualized volatility | 0.2% |
| Deepest drawdown | −0.0% (2021-03-01 to 2021-03-03) |
| Up days | 69.0% |
| Average daily range | 0.01% |
| Average overnight gap | 0.01% |
| Correlation to SPY | -0.01 |
| Correlation to QQQ | -0.01 |
| Correlation to TLT | 0.01 |
| Sessions above the 200-day average | 99.5% |
| Crossings of the 200-day average | 10 |
| Falls of 10% or more from a 20-day high | 0 |
Calendar years
| Year | Return |
|---|---|
| 2021 | 0.1% |
| 2022 | 1.6% |
| 2023 | 5.1% |
| 2024 | 5.3% |
| 2025 | 4.2% |
| 2026 (part) | 2.8% |
Biggest single days
| Best day | Move |
|---|---|
| 2023-04-05 | 0.1% |
| 2023-01-12 | 0.1% |
| 2023-12-14 | 0.1% |
| 2023-12-21 | 0.1% |
| 2023-12-28 | 0.1% |
| Worst day | Move |
|---|---|
| 2021-03-03 | −0.0% |
| 2021-10-12 | −0.0% |
| 2021-08-02 | −0.0% |
| 2021-08-17 | −0.0% |
| 2021-10-14 | −0.0% |
Average return by calendar month
| Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.3% | 0.2% | 0.3% | 0.3% | 0.3% | 0.3% | 0.3% | 0.3% | 0.3% | 0.3% | 0.3% | 0.3% |
Most and least correlated funds
| Most correlated | Least correlated | ||
|---|---|---|---|
| IEI | 0.06 | KMLM | -0.03 |
| UDN | 0.05 | FAS | -0.03 |
| AGG | 0.04 | XLF | -0.03 |
| FXE | 0.04 | TECL | -0.02 |
| IGIB | 0.04 | RINF | -0.02 |
What the SGOV price series looks like
SGOV's calendar returns were 0.05% in 2021, 1.57% in 2022, 5.12% in 2023, 5.29% in 2024, 4.23% in 2025 and 2.76% in 2026 to date. The yearly pattern is the history of short-term rates in the window: almost nothing in 2021, then a steady rise. Up days were 68.95% of sessions, with an average up day of 0.02% and an average down day of -0.01%. The best day was 0.08% on 2023-04-05 and the worst was -0.02%.
The intraday range averaged 0.01% and the average overnight gap was 0.01%. The overnight share of return was 95.25% and the intraday share was 4.75%. The fund earns by holding, between the close and the next open, and almost none of the return is available during the session. A rule that enters at the open and exits during the day has nothing to capture here. Its beta to SPY, QQQ and TLT was 0 in all three cases, and its correlation to SPY was -0.01.
The deepest drawdown was 0.02% from 2021-03-01 to 2021-03-03, recovered on 2021-03-04. The longest drawdown lasted 11 sessions. In a fund like this, drawdown is a rounding effect on the price, and the drawdown column of every template is close to zero until costs are added.
The RSI statistics describe a fund that never gets oversold. RSI(14) was above 70 on 1085 sessions and below 30 on 0. RSI(2) was below 10 on 17 sessions, and the median 5-day and 20-day forward returns after those were both 0.01%, against baselines of 0.07% and 0.32%. A mean-reversion buy signal on this fund came with a lower forward return than any other session. The fund was above its 200-day average on 99.52% of sessions and crossed it 10 times. A fund in a smooth climb stays above the average, so trend rules are always invested and mean-reversion rules are mostly idle.
The monthly averages show a steady accrual, between 0.24% and 0.31% for each calendar month, with 5 or 6 observations each. The day of week averages are 0.01% or 0.02%. Lag-1 autocorrelation was 0.06. There was no pattern in the data beyond the climb.
The funds most correlated with SGOV are IEI at 0.06, UDN at 0.05, AGG at 0.04 and FXE and IGIB at 0.04. Those are all effectively zero. The least correlated, KMLM, FAS and XLF at -0.03, are also zero. SGOV has no relationship to any other fund in the test.
That makes the peers on the bond pages a useful comparison. AGG held at -0.78% and BND at -0.80%, and the best template on both was the RSI(2) snapback at 1.57% and 1.41%. TLT held at -8.23% and IEF at -2.29%, and the RSI(2) snapback made 0.96% and 1.84%. IEI held at -0.36%, and IGIB at 0.21%. On those funds the rules had price movement to work with, and a negative buy-and-hold figure gave a rule that sat out some of it a way to come out ahead. On SGOV the fund's own return was positive and steady, and no rule improved on it.
The note on the fund in the test is that T-bills are often used as the cash leg of rotation strategies. The results here put a number on that choice for this window: 3.22% a year for holding the fund against 0.00% for the same sleeve held as idle cash. The templates that never traded show that 0.00% figure directly.
The limits are the usual ones. One window of 5.74 years, no fees in the headline run, daily-decision rules, and a fund whose returns reflect one rate cycle.
What the overnight split and calendar averages leave open
Over the window the fund's log return split into 17.73% overnight and 0.88% intraday. The average up day was 0.02% and the average down day was -0.01%, so even the largest daily moves in the sample, 0.08% on 2023-04-05 and 0.07% on four other dates in 2023, were small. Those best days cluster in 2023, the year the fund made 5.12%, and the worst days, each -0.02%, cluster in 2021, the year it made 0.05%. The extremes line up with the level of short-term rates in the window.
The calendar-month averages range from 0.24% in February to 0.31% in August and December. With 5 or 6 observations per month, the gap between those figures says nothing about seasonality. It mostly reflects which months fell into the low-rate year of 2021 and which fell into the high-rate years afterwards. A month-of-year rule has nothing to find in this series, which is why the monthly cycle ended behind the fund in every year from 2022 onward.
The day-of-week averages are 0.01% from Monday to Wednesday and 0.02% on Thursday and Friday. A one-point difference of that size is a rounding step in two-decimal data. The lag-1 autocorrelation of 0.06 is also small. A fund like this gives momentum and reversal rules nothing to measure, which agrees with the idle result of the four templates that never traded.
The liquidity figures bound what the test cannot see. The fund trades about $532,191,384 a day, and the median minute carries 8,376 shares. A $10,000 account is tiny against that, so the fills in the backtest are small relative to the volume. The cost runs charge 5 and 10 basis points regardless, which treats SGOV like any other fund in the study. For a T-bill fund with a very tight spread, that is a harsh assumption, and the cost-run losses on this page are better read as a stress test of the templates than as an estimate of what trading SGOV would cost.
Build it from blocks (or type it in English), backtest it on 5.7 years of minute data in seconds, tweak any parameter, then paper trade it on live data. No card, no broker needed to start.
Frequently asked questions
What was the best strategy for SGOV?
Of the 12 templates tested on SGOV over 2021-01-04 to 2026-10-02, the strongest by CAGR was golden cross at 3.2% (max drawdown 0.0%), versus 3.2% for buy-and-hold. The best result in hindsight is not a forecast. Check drawdowns and trade counts before drawing conclusions.
Did any strategy beat buying and holding SGOV?
0 of 12 templates beat SGOV buy-and-hold (3.2% CAGR) on this window; 4 of 12 had a shallower maximum drawdown than holding (0.0%).
What was the best strategy for SGOV?
The golden cross, the 200-day regime filter and the trend plus trailing stop each made 3.22% a year, level with buy-and-hold. They ended between $11,993 and $11,997 against $11,996 for holding. None beat the fund.
Did any strategy beat buying and holding SGOV?
No. Zero of twelve beat the fund's 3.22% CAGR. Eight finished with a positive CAGR. Four templates made no trades and finished at $10,000, a 0.00% return.
Why did four strategies make no trades on SGOV?
RSI(14) mean reversion, the dip buyer, the momentum breakout and the 3-month momentum switch never met their entry conditions. RSI(14) fell below 30 on 0 sessions, and SGOV had 0 falls of 10% or more from a 20-day high.
How much do trading costs matter for SGOV strategies?
A great deal for the active ones. The weekly 7% target went from 3.20% to 1.08% at 5 basis points and -4.85% at 10. The RSI(2) snapback went from 0.02% to -1.73%. Rules with one fill barely changed.
Why is the Sharpe ratio on SGOV so high?
Annualized volatility was 0.24%, so the denominator is tiny. Buy-and-hold shows 13.84 and the golden cross 14.14. The ratio reflects how smooth the accrual was, and several templates with a Sharpe above 13 simply matched the fund.
How volatile is SGOV?
Annualized volatility was 0.24% and the deepest drawdown was 0.02%, recovered in 11 sessions. Up days were 68.95% of sessions. The fund's beta to SPY was 0 and its correlation to SPY was -0.01.
Other bond etfs
Backtests are hypothetical, computed by DeployQuant's engine on minute-resolution consolidated US market data (2021-01-04 to 2026-10-02, $10,000 starting capital, no margin, no fees or slippage in the headline run; buy-and-hold puts 98% of the account in at the first open, as the templates do) and do not guarantee future results. Nothing on this page is investment advice. Live trading involves risk of loss.