A hypothetical backtest of a simple trend rule on TQQQ, QLD, SSO and UPRO since 2017: what it did to returns and worst falls, selloff by selloff and year by year, how often it traded, and where it failed.
In Exent’s backtest it did what an exit rule is for. Holding TQQQ only while the Nasdaq‑100 was above its 200‑day average cut the worst fall from −81.6% to −54.8%, and the return was higher too: 46.4% a year against 42.5% for holding, from Jan 2017 to Sep 2026. The results are hypothetical, past, and before trading costs and taxes. On the S&P 500 funds the same rule also cut losses, but gave up a lot of return.
The rule is simple. Each day, check whether the index closed above its 200‑day moving average the day before. If it did, hold the leveraged fund; if not, hold T‑bills (BIL). The signal uses the index, QQQ for the Nasdaq‑100 funds and SPY for the S&P 500 funds, not the leveraged fund’s own price, which is far more volatile. The average itself is one of the most watched lines in the market: a Zacks article on Yahoo Finance in September 2026 called it one of the most widely used indicators, tracked by algorithms, big funds and retail traders alike.
The same idea on the plain S&P 500 is in does the 200‑day moving average strategy work; the glossary entry tests a monthly version over a different window, so its numbers differ. All the protection tools are compared in the portfolio protection guide.
Data through the close of Oct 19, 2026. The S&P 500 fell 1.5% over the last five trading days, and the Nasdaq‑100 fell 3.7%. The S&P 500 is 1.6% below its record close of Oct 6: a small dip from its record. It is 6.1% above its 200‑day average (+8.0% a week earlier). Its 20‑day volatility is 9.6% a year, calmer than 70% of trading days since 2016.
Daily closes over the last year, through Oct 19, 2026. Green: the index above its average; red: below it.
Year‑to‑date return through Oct 19, 2026: what each fund returned next to its multiple times the index’s return. The gap is the daily reset at work.
| Fund | Last 5 days | Year to date | Multiple × index | From its high |
|---|---|---|---|---|
| 2× S&P 500 (SSO) | −3.1% | +20.8% | +25.4% | −3.7% |
| 3× S&P 500 (UPRO) | −4.6% | +28.4% | +38.1% | −6.0% |
| 2× Nasdaq‑100 (QLD) | −7.5% | +27.5% | +34.2% | −10.9% |
| 3× Nasdaq‑100 (TQQQ) | −11.3% | +34.2% | +51.3% | −19.0% |
| 3× Russell 2000 (TNA) | −7.4% | +16.8% | +31.2% | −50.6% |
| 3× chip stocks (SOXL) | −23.2% | +134.7% | +200.3% | −67.2% |
Last updated Oct 19, 2026. Members only.
| Exent regime pillar | Today |
|---|---|
| Liquidity | |
| Credit stress | |
| Economic cycle | |
| Sentiment | |
| Regime: Risk ON · Neutral · Risk OFF |
Price trend and volatility show what the market has already done. Exent’s macro regime score reads the backdrop behind it (liquidity, credit stress, the economic cycle and sentiment) and classifies it as Risk ON, Neutral or Risk OFF. Market intel members see today’s reading and get an alert when it changes.
It cut every fund’s worst fall by roughly a third or more, and it came close to or beat the holding return only on the Nasdaq‑100 funds. Hypothetical backtest, Jan 2017 – Sep 2026, dividends and T‑bill interest included, no costs or taxes:
| Fund | Hold: a year | Hold: worst fall | 200‑day rule: a year | Rule: worst fall | Time invested |
|---|---|---|---|---|---|
| 2× Nasdaq‑100 (QLD) | 34.3% | −63.7% | 33.2% | −40.2% | 83% |
| 3× Nasdaq‑100 (TQQQ) | 42.5% | −81.6% | 46.4% | −54.8% | 83% |
| 2× S&P 500 (SSO) | 23.6% | −59.3% | 16.2% | −36.7% | 82% |
| 3× S&P 500 (UPRO) | 28.0% | −76.8% | 20.6% | −52.2% | 82% |
The rule was invested about 83% of the time on the Nasdaq side and 82% on the S&P 500 side. The holding returns here start in Jan 2017, once the first 200‑day average exists, so TQQQ’s 42.5% a year differs from the 44.1% from Jul 2016 used in can you hold TQQQ long term.
It helped most in the slow 2022 decline and least in the fastest crash, COVID in 2020. Each row is TQQQ’s largest fall inside the selloff, held and with the rule:
| Selloff | Holding: largest fall | With the 200‑day rule |
|---|---|---|
| Q4 2018 selloff | −57.9% | −41.7% |
| COVID crash, 2020 | −69.9% | −54.8% |
| 2022 bear market | −80.9% | −46.3% |
| 2025 tariff selloff | −56.7% | −26.9% |
The 2022 bear market lasted 195 trading days, which gave the rule time to step aside: TQQQ’s largest fall in that selloff was −46.3% with the rule instead of −80.9%. The COVID crash took 23 trading days. By the time the index closed below its average, much of the damage was already done, and the rule’s worst fall of the whole decade, −54.8%, came there. A price rule cannot react to a fall that has not happened yet.
The rebound years. After each big fall, the index had to climb back above its average before the rule bought back in:
| Year | Holding | With the 200‑day rule |
|---|---|---|
| 2018 | −19.7% | −8.6% |
| 2019 | +134.0% | +58.7% |
| 2020 | +110.1% | +106.0% |
| 2021 | +83.0% | +83.0% |
| 2022 | −79.0% | −43.9% |
| 2023 | +203.0% | +117.3% |
| 2024 | +60.5% | +60.5% |
| 2025 | +35.1% | +35.2% |
| 2026 (to Sep 2026) | +48.2% | +33.0% |
The rule saved a lot in 2018 (−8.6% vs −19.7%) and 2022 (−43.9% vs −79.0%) and gave back a lot in 2019 and 2023, when it missed the first leg of each rebound. In 2025 it ended level with holding (+35.2% vs +35.1%) while cutting the tariff‑selloff fall from −56.7% to −26.9%.
Whipsaws are the other cost: an exit, then a rebound above the average, then a buy back at a higher price. The Nasdaq‑100 signal switched 34 times in nearly ten years and the S&P 500 signal 58 times. On the S&P 500 side 2018 was a whipsaw year: the rule lost more than holding on SPY (−7.8% vs −4.6%), SSO (−18.5% vs −13.9%) and UPRO (−29.1% vs −25.1%), and UPRO’s worst fall with the rule (−52.2%) stretched from Jan 2018 to Jun 2020.
Taxes come on top. In a taxable account every switch out is a sale. Gains on positions held one year or less are short‑term, and the IRS taxes net short‑term gains as ordinary income. Inside a tax‑advantaged account such as an IRA, trades are not taxed when they happen.
Because the deeper the potential loss, the more an exit is worth. Avoiding part of a −33.7% fall spares a fund part of a 50.8% rebound; avoiding part of a −81.6% fall spares it part of a 443.5% rebound. Missed rebounds cost leveraged funds more too, but the losses avoided grew faster than the gains missed. The stretches below the average were also the most volatile: in Exent’s data, all 10 of the S&P 500’s best days and 8 of its 10 worst since mid‑2016 came while it was below its 200‑day average, and that is where daily‑reset funds lose most to decay. Signal noise was the second factor: the S&P 500 signal whipsawed more often.
| Fund | Signal, switches | Holding: worst fall | Gain needed to recover it | Rule’s yearly return as a share of holding’s |
|---|---|---|---|---|
| S&P 500 (SPY) | SPY, 58 | −33.7% | +50.8% | 68% |
| SSO (2×) | SPY, 58 | −59.3% | +145.7% | 69% |
| UPRO (3×) | SPY, 58 | −76.8% | +331.0% | 74% |
| Nasdaq‑100 (QQQ) | QQQ, 34 | −35.1% | +54.1% | 90% |
| QLD (2×) | QQQ, 34 | −63.7% | +175.5% | 97% |
| TQQQ (3×) | QQQ, 34 | −81.6% | +443.5% | 109% |
Within each index, the more leverage, the larger the share of the holding return the rule kept. Only on TQQQ did it come out ahead. The math behind the third column is in loss recovery math.
They answer different questions. A 200‑day rule reacts to price: it can only exit after the index has already fallen through its average. That is why it did least in the fastest crash and whipsawed in sideways years. It is still a rule, written in advance, which is what separates it from selling on fear (see does market timing work).
A macro regime score starts from the conditions behind prices rather than prices themselves. A score of this kind reads broad areas such as liquidity (how much money is flowing through the financial system), credit stress (how strained borrowing and lending are), the economic cycle (whether the economy is expanding or slowing) and sentiment (how fearful or confident investors are). It then classifies the backdrop as risk‑on, neutral or risk‑off. Neither approach is a forecast, and this page tests only the price rule. For how the two fit into a decision about leverage, see when it is safe to hold leveraged ETFs.
In a hypothetical backtest from January 2017 to September 2026, holding TQQQ only while QQQ was above its 200-day average returned 46.4% a year against 42.5% for holding, and cut the worst fall from 81.6% to 54.8%, before costs and taxes.
The backtest uses the index fund, QQQ, not TQQQ’s own price. The index is what the fund multiplies, and its trend is far less noisy than a 3x fund’s price.
On the Nasdaq-100 it switched 34 times in nearly ten years, and on the S&P 500 58 times. In a taxable account every exit is a taxable sale, and none of the results include trading costs or taxes.
The S&P 500’s falls were shallower, so the rebounds the rule missed cost more than the losses it avoided, and its signal whipsawed more often. Holding SPY returned 15.1% a year against 10.3% with the rule.
A rule set in advance that moves a leveraged position to cash or T-bills when a condition is met, such as the index closing below its 200-day average, and moves back when it reverses. It limits the depth of losses but adds trades and missed rebounds.
No. A price rule reacts after prices have moved. A macro regime score reads the conditions behind prices, such as liquidity, credit stress, the economic cycle and sentiment, and classifies them as risk-on, neutral or risk-off.
Market Intelligence reads the market regime and tells you whether the backdrop favors taking risk or protecting capital.
Educational content, not investment advice. Examples are illustrative; past performance does not guarantee future results.