SOXL and leveraged-ETF decay, explained honestly

The short answer

A 3x leveraged ETF like SOXL (3x long semiconductors) or SOXS (3x inverse) aims to deliver three times the daily return of the semiconductor index — not three times the return over a week, a month, or a year. Because it resets that leverage every single day, the maths of compounding works against you whenever the market chops up and down. Over multi-day holds, a 3x ETF can drift below 3x of what you'd expect, and in a choppy or falling market it can bleed value even if the index ends up roughly flat. That erosion is what people mean by "leveraged-ETF decay."

Do single-stock leveraged ETFs decay the same way?

Yes — and usually worse. The 2x single-stock funds (NVDA, TSLA and the rest) run the same daily reset described below, but a single company's daily volatility typically dwarfs an index's, and volatility drag scales with volatility squared. Same mechanism, hotter fuel — plus single-name gap risk (earnings, halts) with no index diversification to soften it. Everything on this page applies to them, amplified.

Why daily reset causes decay

A leveraged ETF promises a multiple of one day's move, then rebalances overnight so it can promise the same multiple again tomorrow. That nightly reset is the whole story. Two days of equal-and-opposite index moves don't cancel out for you the way intuition suggests — they leave you slightly poorer. This is sometimes called volatility decay or beta slippage.

How many times a day does SOXL rebalance? Once. Leverage is reset one time per trading day, with the rebalancing trade executed at or near the 4:00 p.m. ET close, so the 3x multiple starts fresh against the new net asset value at the next open. It is not rebalanced continuously through the session, and there is no second reset.

A worked example

Say the underlying index sits at 100 and a 3x ETF sits at $100. The index drops 10% on day one, then rises 11.11% on day two — landing back exactly at 100, flat over two days.

The index is flat. The 3x ETF is down 6.67%. Nobody made a "mistake" — the product did exactly what it promised each day. The reset just guarantees that a round-trip through volatility costs you. Run that pattern for weeks and the gap compounds. The more the tape whipsaws, the faster the erosion. In a smooth one-direction trend the same compounding can run the other way and add to a move, but markets rarely travel in a straight line — and that upside is not a reason to hold these passively. It is simply the symmetric side of the same mechanism that punishes you in chop.

Volatility is the fuel for decay

Decay scales with how much the underlying bounces around, not with time alone. A calm, steady uptrend erodes a leveraged ETF less; a violent, range-bound market with no net direction is the worst case. Semiconductors are one of the most volatile corners of the equity market, so SOXL and SOXS sit at the high-decay end of the spectrum. That volatility is exactly why some traders are drawn to them — and exactly why holding them passively is so punishing.

SOXL's real record vs "3x the index," year by year

Theory is easy to wave away, so here is the actual arithmetic — every calendar year of SOXL's trading history against the SOXX index ETF and against the naive expectation that SOXL should return three times what SOXX did. Price returns computed from SIP daily closes (first to last trading close of each calendar year, splits adjusted, dividends not reinvested). Download the CSV and recompute it from any clean daily source.

* 2016 starts at SOXL's first SIP session in our archive (June 20); 2026 runs through August 6 — both partial years. † The 2022 "positive gap" is the naive formula breaking, not SOXL winning: 3 × −36.4% implies −109%, and a fund cannot lose more than 100%. Daily reset is precisely what prevents that — which is the honest flip side of decay.

Read the pattern, not any single row. In smooth trending years (2017, 2019, 2023) daily compounding beat naive 3x — that is the mechanism working in your favor. In choppy years it took huge bites: 2020 cost 89 points of naive expectation (the index rose ~50%, SOXL managed +60% instead of +149%), 2024 turned a +17% index year into a losing one, and 2025 gave up 67 points. Same product, same mechanism, opposite outcomes — the difference is purely the path the index took.

The drawdowns that do the real damage

The 2022 episode is the one to sit with: a buyer at the December 2021 peak was down 90.5% at the bottom and waited almost three and a half years to get back to even — through a period in which the underlying index itself recovered far sooner. A −90% hole needs a +900% climb. That asymmetry, not the day-to-day drag, is what ends accounts.

Decay calculator: what should a leveraged ETF return on your scenario?

Enter an index move, a daily volatility, and a holding period. The calculator compares the naive expectation (leverage × index move) against a standard volatility-drag model: (1+R)L × e−(L²−L)/2 · σ²N − 1. It is an approximation — real funds also carry fees, financing costs and tracking error — but it captures the compounding mechanics the yearly table shows.

Precomputed scenarios (same model)

Two scenarios worth staring at: a flat violent quarter quietly costs a 3x fund ~16% — that is pure decay, no market opinion required. And the same +20% index gain returns +66% when the ride is calm but only +36% when it is choppy — the "3x" on the label is a daily promise, and the path decides what you actually keep. For which tickers this applies to on Robinhood, see the single-stock leveraged ETF list.

Gap risk: the part stops can't fully protect

Leverage cuts both ways on overnight and intraday gaps. If the semis index gaps down 10% before you can react, a 3x ETF gaps roughly 30% — before any slippage. A protective stop is an instruction to sell once a price is touched; it does not guarantee that price. In a gap, the market can open well below your stop and fill you far lower. Leverage magnifies every gap, so risk controls that work fine on an unleveraged stock leave a much wider wound here.

Stops and size are the parts of a leveraged trade you actually control. That is the thing Coil sells: a level, not a ticker. Coil Live (MCP feed) serves the scored board to your own AI agent over MCP or plain HTTPS for $29 a month — structural stop, target, entry lane, and the reduced-size rule that rides along with a leveraged vehicle — re-scored about every 5 minutes through the market day. Coil never sees your brokerage account and never places an order; your agent does that at your own broker. The record those rules produced is published every session at /results alongside its benchmark, and it is behind that benchmark right now.

Does SOXL go to zero?

How much would your portfolio drop if SOXL fell 80%? It is position size times the move: a SOXL position worth 5% of your account takes 4% off the account, 10% takes 8%, 25% takes 20%. And an 80% fall is not a hypothetical — SOXL has already printed two drawdowns that deep or worse, −80.6% in the February–March 2020 crash and −90.5% from the December 2021 peak, both in the table above.

An honest answer: a single catastrophic day could in theory wipe out most or all of a 3x long ETF's value (a ~33%+ single-day drop in the underlying implies a ~100% loss before circuit breakers), and prolonged decay plus reverse splits means these funds can trend toward worthlessness over long holds. In practice, providers use reverse splits to keep the share price off zero, and exchange circuit breakers halt extreme single-day moves — so a literal $0 print is rare. But the practical lesson stands: treat total loss of the position as a real, live possibility, not a tail you can ignore. Never hold a 3x ETF as a long-term core position and never size it as if it behaves like the index.

Why they're built for short holds

Put the mechanics together — daily reset, volatility decay, gap-magnified leverage — and the design intent is clear. Leveraged ETFs are short-duration tactical tools. The issuers say so in their own prospectuses: they are intended for investors who actively monitor and manage positions, typically intraday to a few days. Buy-and-hold is the misuse case that generates most of the horror stories.

How a long-only rules engine treats them

This is the lens behind how Coil works, so it's worth a plain note — not a recommendation to trade these instruments, just an explanation of one disciplined approach. Coil is rules-based trading software you run yourself, on your own machine, broker and capital. It is long-only: it uses long leveraged vehicles (like SOXL, NVDL, or TQQQ) to accelerate a leader when the setup earns it, and it never shorts the inverse — when the tape turns down it raises cash and rotates into a defensive Macro book rather than fighting drift with a 3x inverse fund — a choice we measured rather than assumed: we pre-registered a grid of inverse-ETF hedges and measured every cell against cash and a defensive macro sleeve, and all 36 failed. It treats decay as a constraint to design around, not a bug to ignore:

• Reduced size, because leverage is dangerous. Decay and gap-magnified leverage are exactly why the engine sizes leveraged vehicles down — accelerating a leader at reduced notional, never betting the account on a 3x fund.
• Timed entries, not chasing. Positions are opened only at real entry-windows — a name scored READY or SETUP, at support — never on a FALLING knife and never CHASEing a name pinned at its highs, then exited on a laddered scale-out and a trailing stop rather than married across weeks of chop where decay does its damage.
• Standing aside counts. When nothing scores a real entry, the engine holds cash rather than fight the tape. Idle cash can earn the broker's own variable cash sweep (~3.35% APY on Robinhood Gold as of early 2026 — the broker's yield, not Coil's, and not risk-free).
• Risk circuit-breakers tied to the account's high-water mark cap how much a single bad day can do. They reduce damage; they do not remove risk, and stops can still gap through.

The scoring behind those entries is validated, not asserted. In a point-in-time research backtest (2017–2026 H1, survivorship-free with delisted names included, next-open fills, costs modeled), the leadership-rotation backbone Coil is built on compounded +638% versus SPY's +282%, with a shallower worst drawdown (−23% vs −32%) and a positive result in 9 of 10 years (worst −1%, in 2018). The honest rider rides with it: through the end of 2025 it ran roughly even with SPY at about one-third less drawdown — the outperformance concentrates in leadership regimes (2025 +51%, 2026 H1 +86%). These are research figures, not live results; the engine is newly live, and none of this makes leveraged ETFs "safe." We publish that live record free and unauthenticated at /api/perf — engine versus SPY and QQQ since the engine's 2026-07-08 inception, funding-adjusted, percentages only, and omitted rather than estimated when the data will not support it. If you want to weigh tooling approaches, see Coil vs. trading bots and signal services, or read the engine details on the pricing and download page.

FAQ

What does it mean that SOXL resets every day?

It means SOXL targets 3x the return of one trading day — not 3x the return over a week, a month, or a year. Each night the fund rebalances so it can promise the same 3x multiple again tomorrow. That nightly reset is why multi-day results drift from 3x: two equal-and-opposite index moves don't cancel out. If the index falls 10% then rises 11.11%, it ends flat — but a 3x ETF goes $100 to $70 to $93.33, down 6.67%. The product did exactly what it promised each day; the reset simply guarantees that a round-trip through volatility costs you.

Can SOXS or SOXL go to zero?

An honest answer: a single catastrophic day could in theory wipe out most or all of a 3x ETF's value — a roughly 33% adverse single-day move in the underlying implies a near-total loss before circuit breakers. In practice a literal $0 print is rare, because issuers use reverse splits to keep the share price off zero and exchange circuit breakers halt extreme single-day moves. But prolonged decay plus repeated reverse splits means these funds can trend toward worthlessness over long holds. Treat total loss of the position as a real, live possibility rather than a tail you can ignore.

Are SOXL and SOXS worth holding overnight?

Overnight is where leverage is least forgiving. If the semiconductor index gaps down 10% before you can react, a 3x ETF gaps roughly 30%, before any slippage. A protective stop is an instruction to sell once a price is touched — it does not guarantee that price. In a gap the market can open well below your stop and fill you far lower. Leverage magnifies every gap, so risk controls that work fine on an unleveraged stock leave a much wider wound here. This is mechanical risk, not a prediction about any particular night.

What happens if you hold SOXL long term?

Decay scales with how much the underlying bounces around, not with time alone. A calm one-direction uptrend erodes a leveraged ETF relatively little and compounding can even work in your favour; a violent, range-bound market with no net direction is the worst case. Semiconductors are among the most volatile corners of the equity market, so SOXL and SOXS sit at the high-decay end of the spectrum. Issuers state in their own prospectuses that these products are intended for investors who actively monitor and manage positions, typically intraday to a few days. Buy-and-hold is the misuse case behind most leveraged-ETF horror stories.

Can SOXL wipe out?

A total single-day wipeout is unlikely — circuit breakers halt extreme index moves and issuers use reverse splits to keep share prices off the floor — but one catastrophic session can erase most of a 3x position, and prolonged chop can decay it toward worthlessness over long holds. Treat total loss as a real possibility, not a technicality.

Has SOXL ever lost money in a year the index went up?

Yes. In 2024 the SOXX semiconductor index rose 17.1% while SOXL fell 1.8% — about 53 percentage points short of the naive 3x expectation, caused entirely by daily-reset compounding through a choppy year. 2020 is the starker case: the index rose 49.6%, naive 3x implies +149%, but SOXL delivered +60% because the March crash and recovery burned roughly 89 points of compounding drag. Both are price returns computed from SIP daily closes; the full 2016–2026 table is published on this page with a downloadable CSV.