TMF observed max drawdown: 92.3%
Direxion 20+ Year Treasury Bull 3x. From its 2020-07-31 peak, TMF fell 92.3% to the 2026-07-31 trough and has not yet made a new high. Computed from 209 months of observed fund history. Fund inception boundary (first available fund observation): 2009-04-16; the monthly analysis runs from 2009-04-30 through 2026-08-31.
Before 2009-04-16, the extended history is simulated. Observed fund drawdowns come first; simulation assumptions and results are separated below.
What is TMF's observed maximum drawdown?
TMF's observed maximum drawdown was 92.3%, from a peak in 2020-07-31 to a trough in 2026-07-31, based on monthly closes from fund prices first observed on 2009-04-16. The sample contains 209 monthly observations through 2026-08-31. Monthly closes can miss deeper daily losses.
| Measure | Value |
|---|---|
| Maximum drawdown | -92.3% |
| Peak | 2020-07-31 |
| Trough | 2026-07-31 |
| Recovered | Not yet recovered |
| Monthly observations | 209 |
Last verified
What is TMF's observed maximum drawdown?
TMF's observed maximum drawdown was 92.3%, from a peak in 2020-07-31 to a trough in 2026-07-31, based on monthly closes from fund prices first observed on 2009-04-16.
How long did TMF take to recover from its worst drawdown?
TMF has not yet recovered from its worst observed drawdown, which began at the 2020-07-31 peak.
What was TMF's worst 12-month return?
TMF's worst observed rolling 12-month return was -75.1%, in the 12 months ending 2022-10-31.
TMF observed fund drawdown statistics
2020-07-31 peak to 2026-07-31 trough.
Still below the prior peak.
Share of all months spent below a prior high.
Rolling year ending 2022-10-31.
The biggest TMF drawdowns in observed fund history
| Depth | Peak | Trough | Recovered | Peak to trough | Total underwater |
|---|---|---|---|---|---|
| -92.3% | 2020-07-31 | 2026-07-31 | not yet | 72 mo | 6.1 years |
| -49.7% | 2012-07-31 | 2013-12-31 | 2015-01-31 | 17 mo | 2.5 years |
| -47.0% | 2016-07-31 | 2018-10-31 | 2019-08-31 | 27 mo | 3.1 years |
| -40.8% | 2010-08-31 | 2011-01-31 | 2011-09-30 | 5 mo | 13 months |
| -39.1% | 2015-01-31 | 2015-06-30 | 2016-06-30 | 5 mo | 17 months |
Methodology: observed TMF fund prices only, first available from 2009-04-16. Monthly closing-price statistics cover 2009-04-30 through 2026-08-31; no proxy or pre-inception simulation is included in this table. Educational information, not investment advice.
Before 2009-04-16: simulated from the underlying index with assumptions
This separate series combines observed prices with a modeled pre-inception extension from 1970-01-31 through 2026-09-30. The 2026-09-30 endpoint is the legacy current partial-month bucket/month-end label, not evidence that TMF was observed through that future date. It is not observed TMF performance. The API's observed boundary is 2009-04-16; observed monthly statistics run from 2009-04-30 through 2026-08-31.
- Rebalancing convention: monthly rebalance at month-end close; drawdown statistics use monthly closing prices.
- Leverage factor: 3x daily returns of TLT, the underlying index proxy, with daily leverage reset and compounding.
- Expense proxy: 1.09% annual expense ratio, deducted daily using 252 trading days per year.
- Financing proxy: FRED DFF (Federal Funds Rate), forward-filled to return dates, with a 0.50% spread on financed exposure (leverage factor minus one); borrowing pays the rate plus spread, while lending earns the rate minus spread.
- When FRED DFF is unavailable, the rate proxy is 2.50%, with the same spread applied.
- Proxy source for TMF: TLT (synthetic_leverage), from 1920-02-03 to 2009-04-15.
- Proxy source for TLT: VUSTX (proxy_etf), from 1986-05-19 to 2002-07-25.
- Proxy source for TLT: us_30yr (static_monthly), from 1920-01-31 to 1986-05-16.
- Proxy source for SPY: VFINX (proxy_etf), from 1976-08-31 to 1993-01-28.
- Proxy source for SPY: spy_tr (static_monthly), from 1900-01-31 to 1976-08-30.
Modeled extended statistics
2020-07-31 peak to 2026-09-30 trough.
Still below the prior peak.
Share of all months spent below a prior high.
Rolling year ending 2022-10-31.
For context, SPY (the S&P 500 baseline) had a maximum drawdown of 50.8% over the same kind of monthly-close analysis.
Modeled extended drawdowns
| Depth | Peak | Trough | Recovered | Peak to trough | Total underwater |
|---|---|---|---|---|---|
| -92.6% | 2020-07-31 | 2026-09-30 | not yet | 74 mo | 6.2 years |
| -90.1% | 1972-12-31 | 1981-09-30 | 1986-03-31 | 105 mo | 13.3 years |
| -66.3% | 1986-08-31 | 1990-04-30 | 1993-08-31 | 44 mo | 7.0 years |
| -61.9% | 2008-12-31 | 2010-03-31 | 2012-05-31 | 15 mo | 3.4 years |
| -49.7% | 2012-07-31 | 2013-12-31 | 2015-01-31 | 17 mo | 2.5 years |
Monthly closing prices, peak-to-trough on total return where available. The pre-inception segment is a reconstruction, not a live fund record, and actual tracking, financing, taxes, spreads, and volatility drag can differ.
Drawdowns are the reason tactical strategies exist
Buy-and-hold TMF means living through every number on this page. BestFolio tracks published TAA strategies whose whole job is cutting these drawdowns, with out-of-sample track records.