DeployQuantBND trading strategies, backtested
Vanguard Total Bond Market ETF: Vanguard's total US bond market fund. Every DeployQuant template run on BND over 5.7 years of minute data, same engine, same window, sorted by return.
BND is the Vanguard Total Bond Market ETF. The 12 DeployQuant templates ran on it from 2021-01-04 to 2026-10-02, each starting with $10,000. Buying BND on the first day and holding it ended at $9,549, a 0.8% annual loss, with a 17.64% maximum drawdown and a Sharpe ratio of negative 0.12. The drawdown began on 2021-01-04, reached its low on 2022-10-20 and had not recovered when the data ended. The fund never regained its first-day level.
That makes BND a different test from the stock funds. Nine of the 12 templates beat buy-and-hold on return, 10 had a shallower drawdown, and 5 finished with a positive return. The best was the RSI(2) snapback at 1.41% a year, which ended at $10,836. The gains are small in dollar terms, and the ranking is mostly a ranking of which rules lost the least in the 2022 bond decline.
Two templates made no trades at all. The dip buyer and the momentum breakout ended at exactly $10,000, which counts as beating a fund that lost money. The window is 5.7 years, and it includes a 13.11% fall for the fund in 2022.
| Strategy | CAGR | max DD | Sharpe | trades | win rate | final value |
|---|---|---|---|---|---|---|
| RSI(2) Dip Snapback | 1.4% | −9.1% | 0.39 | 172 | 60% | $10,836 |
| EMA 12/26 Trend | 1.0% | −5.0% | 0.29 | 22 | 36% | $10,602 |
| Golden Cross (SMA 50/200) | 0.7% | −6.9% | 0.21 | 5 | 40% | $10,422 |
| SMA 10/50 Trend | 0.5% | −6.6% | 0.16 | 19 | 42% | $10,305 |
| 200-Day SMA Regime Filter | 0.4% | −6.5% | 0.12 | 21 | 29% | $10,213 |
| 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 |
| RSI(14) Mean Reversion | −0.2% | −11.6% | -0.03 | 16 | 69% (+1 open) | $9,870 |
| Weekly Entry + 7% Target | −0.5% | −15.5% | -0.06 | 67 | 6% (+1 open) | $9,733 |
| First-to-Last Day of Month | −0.9% | −18.9% | -0.15 | 69 | 54% (+1 open) | $9,465 |
| 3-Month Momentum Switch | −1.0% | −6.5% | -0.55 | 3 | 0% | $9,450 |
| SMA-200 Trend + 15% Trailing Stop | −1.1% | −18.7% | -0.19 | 1 | 0% (+1 open) | $9,395 |
| Buy & hold BND | −0.8% | −17.6% | -0.12 | – | – | $9,549 |
The ranking
The RSI(2) snapback leads at 1.41% with a 9.14% drawdown and a Sharpe ratio of 0.39. It made 172 round trips, won 60% of them and held for an average of 5.1 days, with exposure of 42.7%. The average winner was 0.41% and the average loser 0.49%, so each trade earned or lost less than a half percent. The profit factor was 1.25. Its worst trade, a 4.46% loss over 30 days from 2022-08-30 to 2022-09-29, came in the worst month of the bond decline.
Next are the trend rules. The EMA 12/26 trend returned 1.02% with a 4.96% drawdown, the shallowest of any template that traded. It made 22 round trips, won 8, and had an average winner of 2.06% against an average loser of 0.73%. Its best trade was 5.03% over 152 days from 2024-05-10 to 2024-10-09. The golden cross returned 0.72% from 5 round trips, with two wins averaging 4.2% and losses averaging 1.31%. The SMA 10/50 trend returned 0.52% and the 200-day regime filter 0.37%. The 200-day filter made 21 round trips and won 6, a 29% win rate, and its two long holds, 407 days from 2023-11-28 and 356 days from 2025-05-27, made 5.09% and 4.57%.
Exposure shows how differently the rules used the fund. The weekly target held a position 90.5% of the time and the monthly cycle 95.2%, and both ended below holding or barely above it. The golden cross held 57.5% of the time, the 200-day filter 55.5%, the EMA 12/26 trend 53%, the SMA 10/50 trend 51.4%, the RSI(14) rule 44.1% and the RSI(2) snapback 42.7%. The rules that spent about half their time out of the fund did best, which is what a falling fund rewards. The 3-month momentum switch spent the least time in the fund of any template that traded, at 10.7%, and still lost money, because each of its three entries lost.
The losing group starts with the RSI(14) mean reversion at negative 0.23%. It won 11 of 16 round trips, a 69% win rate, and still lost money because its average loser of 3.25% was larger than its average winner of 1.65%. A single trade opened on 2022-01-04 and held 143 days lost 7.63%. The weekly 7% target returned negative 0.47%. It made 67 round trips and won 4, a 6% win rate, with an average winner of 7.05% and an average loser of 0.47%. The 7% target is far from the fund's normal range, since the average daily move was 0.28%, so the rule mostly closed small losses. Its drawdown was 15.53%, close to holding's 17.64%.
The monthly cycle returned negative 0.95% and lost to holding by a small margin. It held a position 95.2% of the time and its 18.92% drawdown was deeper than the holder's 17.64%. The 3-month momentum switch returned negative 0.98% from 3 round trips, all losers, with exposure of 10.7%. The trend plus trailing stop returned negative 1.08%, from a single closed trade that lost 14.26% over 333 days from 2021-10-28 to 2022-09-26. A second position opened on 2023-01-10 at $63.94 and was 9.37% above entry at the last close.
The two templates with no trades did so because their entry conditions never fired. The dip buyer buys after a drawdown worse than 10% within 20 sessions, and BND had no such fall in this window. The momentum breakout buys after a gain above 10% in 20 sessions, and BND never gained that much that quickly. Both ended at $10,000 with a Sharpe ratio of zero. They beat a falling fund by staying in cash and say nothing about skill.
The 2022 bond decline
BND lost 13.11% in calendar 2022, and holding the sleeve lost 12.8%. The 2022 drawdown ran from 2021-01-04 to 2022-10-20 and the fund had no recovery by 2026-10-02. In that year the templates split by exposure. The EMA 12/26 trend gained 0.4%, a 13.2 point gap in its favour. The 200-day filter lost 0.4%, the golden cross 1.5%, the SMA 10/50 trend 2.1% and the RSI(2) snapback 5.7%. The weekly target lost 7.9%, the RSI(14) rule 8.8%, and the monthly cycle 13.8%, which was worse than holding. The trailing stop strategy lost 13.8% from its position opened in 2021 and closed on 2022-09-26.
Rules that moved to cash early did best. The EMA 12/26 and the 200-day filter react to a falling price within weeks. BND's best day of the window, 2.06% on 2022-11-10, came shortly after the low. Its worst days, negative 1.62% on 2022-06-13 and negative 1.25% on 2022-09-26, were small by stock standards. The damage came from small moves adding up: the average daily move was 0.28%, and the fund still lost 13.11% in the year.
In 2023, 2024 and 2025 the picture turned. BND returned 5.5% in 2023, 1.3% in 2024 and 6.9% in 2025 on a holding basis, and the templates trailed in most of those years. The EMA 12/26 trend trailed by 0.1 points in 2023, 1.8 in 2024 and 3.8 in 2025. The 200-day filter trailed by 5.9, 0 and 3.2 points. The RSI(2) snapback gained 7.2% in 2023, beating holding by 1.7 points, then trailed in 2024 and 2025. A rule that exits a bond fund after a fall can lag the recovery, as it does in stocks.
The year-by-year pattern gives the beat-holding count its meaning. Most templates beat holding because of 2022, and gave back part of that margin in the next three years. Over the whole window the margin was small.
How each strategy traded BND
| Strategy | Time in market | Avg hold (days) | Best trade | Worst trade | Profit factor | With 10 bps slippage |
|---|---|---|---|---|---|---|
| RSI(2) snapback | 42.7% | 5 | 1.5% | −4.5% | 1.25 | −4.3% |
| EMA 12/26 trend | 53.0% | 51 | 5.0% | −2.0% | 1.58 | 0.3% |
| golden cross | 57.5% | 243 | 4.4% | −2.4% | 2.11 | 0.6% |
| SMA 10/50 trend | 51.4% | 57 | 4.6% | −1.9% | 1.28 | −0.1% |
| 200-day regime filter | 55.5% | 56 | 5.1% | −1.6% | 1.27 | −0.3% |
| dip buyer | 0.0% | – | – | – | – | 0.0% |
| momentum breakout | 0.0% | – | – | – | – | 0.0% |
| RSI mean reversion | 44.1% | 53 | 3.4% | −7.6% | 1.08 | −0.8% |
| weekly 7% target | 90.5% | 27 | 7.2% | −1.6% | 0.91 | −3.0% |
| monthly cycle | 95.2% | 28 | 4.4% | −3.3% | 0.87 | −3.3% |
| 3-month momentum | 10.7% | 76 | −0.9% | −3.2% | – | −1.1% |
| trend + trailing stop | 80.6% | 333 | −14.3% | −14.3% | – | −1.1% |
Costs and the busy rules
Slippage matters more for a bond fund than for a stock fund, because the per-trade edge is measured in tenths of a percent. The RSI(2) snapback, with 344 fills, went from 1.41% a year to negative 1.51% at 5 basis points and negative 4.35% at 10. Its drawdown went from 9.14% to 12.91% and 22.79%, and its Sharpe ratio from 0.39 to negative 0.384 and negative 1.14. At 10 basis points the rule ended at $7,746 from a start of $10,000, against $9,549 for holding.
The weekly target went from negative 0.47% to negative 1.61% and negative 2.99%, with a drawdown of 23.34% at 10 basis points. The monthly cycle went from negative 0.95% to negative 2.11% and negative 3.26%. The trend rules lost their edge more slowly. The EMA 12/26 trend went from 1.02% to 0.65% and 0.27%. The SMA 10/50 trend went from 0.52% to 0.2% and negative 0.12%, and the 200-day filter from 0.37% to 0.01% and negative 0.35%. The golden cross held up best, from 0.72% to 0.64% and 0.56%, because it made 10 fills in 5.7 years.
This is the clearest cost result on this page. A rule whose average trade is worth 0.41% cannot absorb 10 basis points on each side. The ranking at 10 basis points is the golden cross, then the EMA 12/26 trend, with the RSI(2) snapback near the bottom. The two no-trade templates are unaffected and finish at $10,000.
BND's average daily dollar volume was $442,813,583 and its median minute volume was 8,357 shares. The costs here are slippage only. Real fills in a bond fund depend on the spread.
What individual trades did on BND
The slow rules on BND made so few trades that single positions decide the result. The golden cross had 5 round trips and 2 winners, with an average hold of 242.6 days. The first winner ran 418 days from 2024-01-04 at 65.66 to 2025-02-25 at 68.55 and returned 4.4%. The second ran 509 days from 2025-03-05 and returned 3.99%. The losers were a 196-day hold from 2023-02-07 that lost 2.39% and an 84-day hold from 2021-10-19 that lost 1.33%. The profit factor of 2.11 is two large wins against two small losses and one scratch, a sample too small to carry much weight.
The trend plus trailing stop rule is one trade and one open position. It bought on 2021-10-28 at 72.22 and sold on 2022-09-26 at 61.92 for a 14.26% loss after 333 days. That exit came shortly before the fund's low of 2022-10-20, so the stop did not save the position, it sold most of the way down. The rule bought again on 2023-01-10 at 63.94 and was still holding at the end of the window with a 9.37% gain. Its 18.68% drawdown is close to the 18.09% drawdown of the fund itself, and it was invested 80.6% of the time.
The 3-month momentum rule lost on all 3 trades: 3.2% from 2023-01-12, 1.57% from 2023-12-21 and 0.94% from 2024-08-02. Each lasted between 49 and 90 days and the median return was negative 1.57%. It was in the market 10.7% of the time and finished with a CAGR of negative 0.98%. Each entry came after a rally that did not continue.
The high-count rules show the opposite problem. The weekly 7% target made 67 round trips and only 4 won, a win rate of 6%. The average win was 7.05% and the average loss 0.47%, which gives a profit factor of 0.91, so the rare large winner did not cover the steady small losses. It held for 27.1 days on average and was invested 90.5% of the time. The RSI mean reversion rule had a better hit rate, 11 winners from 16 trades, yet its average win of 1.65% was smaller than its average loss of 3.25%. Its worst trade lost 7.63%, which is enough to wipe out several winners.
The RSI(2) snapback is the other pattern: 172 trips, a 60% win rate, an average win of 0.41% and an average loss of 0.49%, held for 5.1 days. The profit factor of 1.25 is thin, and the cost runs above show how little room that leaves. The test stops at one 5.74-year window with one parameter set per rule, so none of these trade-level results say how the rules would behave in a different rate cycle.
How BND behaved
| Measure | BND |
|---|---|
| Data in this test | 2021-01-04 to 2026-10-02 (1444 sessions) |
| Total return, buy and hold | −4.7% |
| Annualized volatility | 5.8% |
| Deepest drawdown | −18.1% (2021-01-04 to 2022-10-20) |
| Up days | 48.9% |
| Average daily range | 0.34% |
| Average overnight gap | 0.20% |
| Correlation to SPY | 0.21 |
| Correlation to QQQ | 0.21 |
| Correlation to TLT | 0.92 |
| Sessions above the 200-day average | 64.3% |
| Crossings of the 200-day average | 42 |
| Falls of 10% or more from a 20-day high | 0 |
Calendar years
| Year | Return |
|---|---|
| 2021 | −1.7% |
| 2022 | −13.1% |
| 2023 | 5.7% |
| 2024 | 1.4% |
| 2025 | 7.1% |
| 2026 (part) | −2.7% |
Biggest single days
| Best day | Move |
|---|---|
| 2022-11-10 | 2.1% |
| 2022-09-28 | 1.6% |
| 2023-12-13 | 1.3% |
| 2023-03-10 | 1.2% |
| 2023-11-14 | 1.2% |
| Worst day | Move |
|---|---|
| 2022-06-13 | −1.6% |
| 2022-09-26 | −1.3% |
| 2022-03-02 | −1.2% |
| 2025-04-07 | −1.2% |
| 2024-04-10 | −1.2% |
Average return by calendar month
| Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.4% | −0.5% | −0.4% | −0.7% | 0.2% | 0.3% | 0.7% | −0.1% | −1.3% | −0.9% | 2.0% | 0.1% |
Most and least correlated funds
| Most correlated | Least correlated | ||
|---|---|---|---|
| AGG | 1.00 | TBF | -0.92 |
| IEF | 0.98 | KMLM | -0.44 |
| IGIB | 0.96 | USDU | -0.39 |
| IEI | 0.94 | CTA | -0.36 |
| TLT | 0.92 | RINF | -0.35 |
What BND's own numbers say
BND returned negative 4.67% over 1444 sessions with annualized volatility of 5.78%. Its deepest drawdown was 18.09% and the longest drawdown lasted 1443 sessions, the entire window. The calendar years read negative 1.73% in 2021, negative 13.11% in 2022, 5.68% in 2023, 1.37% in 2024, 7.06% in 2025 and negative 2.67% for 2026 so far. The fund was up on 48.93% of days, the average up day was 0.28% and the average down day 0.29%. The average daily range was 0.34% and the average overnight gap 0.2%.
The fund had no fall of 10% or more from a 20-day high, which is why the dip buyer never traded. Daily moves were small, so the rules built around large moves, the 7% weekly target, the 10% breakout and the 8% dip target, found nothing to act on.
The RSI evidence shows why the mean-reversion rules struggled. The 2-period RSI fell below 10 on 176 sessions. The median 5-day forward return after those sessions was 0%, against a baseline median of negative 0.01%, and the median 20-day return was 0.1% against 0.07%. The 14-period RSI fell below 30 on 57 sessions, with a median 5-day return of negative 0.05% and a median 20-day return of negative 0.85%. The RSI(14) rule bought into falling prices that kept falling. The RSI(2) rule found a small edge of about a tenth of a percent per trade, and the cost runs show that disappears at 5 basis points.
BND closed above its 200-day average on 64.34% of sessions and crossed the line 42 times. The count explains why the 200-day filter made 21 round trips with a 29% win rate. A slow-moving fund hovers around its average, and each crossing is a new trade.
The fund's return came from different parts of the day. Its overnight log return was 4.42% and its intraday log return negative 9.19%, so the fund gained between the close and the open and lost during the session. The 8,357-share median minute volume is large enough for small orders.
Correlation tells the rest. BND's correlation to AGG was 1.0 and to TLT 0.92, to SPY 0.21 and to QQQ 0.21. For comparison, AGG held for negative 0.78% a year and its best template, the RSI(2) snapback, returned 1.57%. IEF held for negative 2.29% with a best template of 1.84%, TLT held for negative 8.23% with 0.96%, and IGIB held for 0.21% with a best template of 2.35%. SGOV, a 0 to 3 month Treasury fund, returned 3.22%, higher than every BND template. Month averages are thin: November averaged 2.05% and September negative 1.29%, from 5 or 6 observations each.
The limits are the same as the rest of the set: one window, no fees in the headline runs, daily decisions, and a bond fund that fell for the first two years and then moved sideways.
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 BND?
Of the 12 templates tested on BND over 2021-01-04 to 2026-10-02, the strongest by CAGR was RSI(2) snapback at 1.4% (max drawdown 9.1%), versus −0.8% 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 BND?
9 of 12 templates beat BND buy-and-hold (−0.8% CAGR) on this window; 10 of 12 had a shallower maximum drawdown than holding (17.6%).
Which strategy worked best on BND?
The RSI(2) snapback returned 1.41% a year and ended at $10,836 from $10,000, with a 9.14% drawdown. The EMA 12/26 trend returned 1.02% with a 4.96% drawdown. Holding BND returned negative 0.8%.
Why did two templates end at exactly $10,000 on BND?
The dip buyer and the momentum breakout made no trades. The dip buyer needs a 10% fall within 20 sessions and the breakout needs a 10% gain within 20 sessions, and BND did neither in this window.
How many templates beat buy-and-hold on BND?
Nine of 12 beat holding on return and 10 had a shallower drawdown. Holding lost 0.8% a year with a 17.64% drawdown. Only 5 templates finished with a positive return.
What happened to BND strategies in 2022?
BND lost 13.11% in 2022. The EMA 12/26 trend gained 0.4% and the 200-day filter lost 0.4%. The monthly cycle lost 13.8% and was worse than holding.
Do trading costs matter on BND?
Yes, because the per-trade edge is small. At 10 basis points per fill the RSI(2) snapback fell from 1.41% to negative 4.35% a year, and the weekly target from negative 0.47% to negative 2.99%.
Did any strategy beat cash on BND?
SGOV, a 0 to 3 month Treasury fund, returned 3.22% in the same window, more than any BND template. The best BND template returned 1.41%.
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.