The Missing 40

Ask a financial adviser anywhere in the world for a sensible portfolio and you will hear the same recipe: 60% stocks, 40% bonds. The 60 does the growing. The 40 keeps the boat steady—which is why finance borrowed a sailing word for it: ballast. On a ship, ballast is the heavy material—stone, iron, seawater—loaded low into the hull. It does not make the ship faster. It keeps the ship upright when the weather turns. In a portfolio, ballast is an asset that holds its value while the growth assets are being thrown around, so the whole portfolio never sinks as far as its riskiest part. That recipe, the 60/40, has been the benchmark for balanced investing for roughly fifty years.

Crypto copied the 60 and never found a 40. BTC and ETH turned out to be two of the great growth assets of the era, but everything tried as ballast beside them has failed the audition: the major alternative coins move nearly in step with BTC, and stablecoins hold their price by earning nothing at all. A crypto portfolio has been all engine and no ballast since the day it existed. So we measured what the last six years would have looked like if crypto had a real 40—built once on BTC and once on ETH, because a portfolio should be judged on the asset you actually hold.

First, the yardstick

Two of the three numbers above are Sharpe ratios, so the term should earn its keep before anything else does. The Sharpe ratio answers one question: how much did an investment earn above cash, per unit of volatility—how widely its daily returns swing, scaled to a yearly figure? Divide the return above cash by the swing and you get a score on which a wild asset and a calm one can compete fairly. Higher is better; zero means cash did just as well; a negative score means cash did better. We lean on it for the same reason the rest of the industry does: raw return flatters whatever happened to swing hardest in a lucky direction, while return per unit of risk measures what an investor was actually paid for what they had to sit through—the only fair scoreboard when one asset swings four times harder than another.

The measure has a history. William F. Sharpe proposed it in 1966 under the name reward-to-variability ratio; the industry kept the idea, shortened the name to his, and has used it as the default scorecard for fund performance ever since. The work earned him a share of the 1990 Nobel Memorial Prize in Economic Sciences. Every Sharpe ratio in this article is computed the same way: daily returns, with cash set to the 3-month US Treasury yield.

How we measured it

The token numbers come from our own pricing engine: daily returns for RiskOFF and for the BTC and ETH it is built on, from the last day of 2019 through July 2026¹. The benchmarks—the S&P 500, US bonds, gold, and Treasury bills—come from the standard public sources². Two portfolios do the work: a BTC 60/40, which is 60% BTC and 40% RiskOFF BTC³, and an ETH 60/40, which is 60% ETH and 40% RiskOFF ETH. Each is rebalanced monthly, and each faces the same judges: its own raw asset, and the classic 60/40 of stocks and bonds run the same way. Nothing in any portfolio forecasts anything. The only decision ever taken is a calendar entry.

The job description of a 40

A 40 has three duties. First, it must hold its value in the storm—when the growth side falls by two-thirds, the ballast is the part that does not. Second, it should earn something while it waits: ballast that pays nothing is dead weight, and dead weight quietly loses to everything around it. Third, it must still be worth something on the day the growth side goes on sale. Rebalancing—selling whichever side has drifted above its target weight and buying the one that has fallen below—is how a portfolio turns a crash into a purchase, and it only works if the 40 kept its value while the 60 was falling. A ballast that crashes with the engine has nothing left to spend at the bottom.

The recipe is older than its mathematics. Balanced stock-and-bond funds date to 1929, when the Wellington Fund carried a bond sleeve through the Depression; Harry Markowitz supplied the formal argument in 1952, in the paper that founded modern portfolio theory and later shared the same 1990 Nobel that honored Sharpe. Somewhere between those two dates and now, the institutional world settled on 60/40 as the reference blend, and it has kept the job ever since.

The year the ballast sank

Then 2022 happened to it. The classic 60/40 lost 15.8%—not because stocks fell 18%, which is what stocks do, but because bonds fell 13% in the same year, the one thing the recipe assumes cannot happen. By one long-running tally it was the worst year for US bonds since 1937⁴, and the worst year for the blended portfolio since the global financial crisis.

And 2022 was not the whole problem. Across our full window, US bonds returned 0.8% a year while cash—in the TradFi sense, meaning 3-month Treasury bills—paid 2.9%. On the yardstick above, that is a Sharpe ratio of −0.30. For six straight years, the world’s default ballast was a worse place to stand than a money-market account.

Crypto never had one

Crypto’s problem is the mirror image. It has never lacked growth. It has lacked anything worth putting beside the growth, and the shortage shows up in the oldest statistic in this article: correlation, the measure Karl Pearson formalized in the 1890s for how two series move together. It runs from +1 (lockstep) through 0 (strangers) to −1 (opposites). Ballast lives near zero or below: an asset that ignores the engine’s weather. Every candidate crypto has produced lives near +1 instead. Over the last six years the major alternative coins’ daily moves ran 0.42 to 0.82 correlated with BTC’s—statistically the same trade wearing different logos. And the coins at the low end of that range are not calm—they are chaotic. DOGE, the least correlated of the majors, lost 40% in a single day inside the window; SOL lost 42% in one. Low correlation earned through wildness is not ballast. It is a second storm.

The other candidate is stablecoins, and they fail the second duty instead. A stablecoin holds its dollar price by holding nothing else: it pays no yield, so it does not even keep pace with the T-bills that TradFi calls cash—the same 2.9% a year that made the classic 40’s six years look so poor. Parking the 40 in a stablecoin is not ballast either. It is an anchor made of paper.

Rebuilt with a real 40

RiskOFF is our candidate for the missing 40: a token built on BTC or ETH that gives up most of its asset’s upside in exchange for a floor under each epoch’s losses. It holds its value in crypto’s storms, and it earned while it waited—11.9% a year on ETH and 8.2% on BTC across the window. Whether it fills the job is exactly what the two portfolios test.

So we put it to work. Sixty cents of every dollar goes to the asset; forty cents goes to the RiskOFF built on that asset; once a month, the portfolio is traded back to 60/40⁵. Over six and a half years of data that meant 78 rebalances per portfolio. On the ETH version, 42 of them trimmed the engine after it ran and 36 added to it after it fell, and the median trade moved just 2.8% of the portfolio; the BTC version split 45 and 33 with a median of 1.9%. Small, boring, mechanical—and that is the whole strategy.

Two more yardsticks appear in the table, both defined here. CAGR, the compound annual growth rate, is the steady yearly rate that turns the first dollar into the last. The Calmar ratio divides that growth by the deepest drawdown—the fall from a peak to the lowest point after it—so it scores growth per unit of worst pain. It is the youngest measure here: Terry W. Young introduced it in 1991 and named it after his newsletter, the CALifornia Managed Account Reports. Futures traders adopted it because leverage makes the deepest fall the number that ends careers, and it stuck as the industry’s pain-adjusted score.

Six and a half years of daily data, January 2020 through July 2026. CAGR and volatility are per year; cash for the Sharpe ratio is the 3-month US T-bill (mean 2.9%). Crypto rows use all calendar days; stock-and-bond rows use NYSE trading days.

The ETH 60/40 (60% ETH, 40% RiskOFF ETH) kept essentially the entire risk-adjusted score of the era’s best asset—0.87 against ETH’s 0.88—while the deepest fall shrank from −79% to −59% and the days worse than −5% fell from 177 to 78. The BTC 60/40 (60% BTC, 40% RiskOFF BTC) paid a visible toll—0.78 against BTC’s 0.81—in exchange for fifteen points less drawdown and 35 ugly days instead of 93. One got the safety free; one paid a fair price for it. Both beat the classic 60/40’s 0.56 by a distance.

A dial, not a trade-off

Forty percent is a convention, not a law. The ballast weight can sit anywhere from 0 to 100, and every setting is a different portfolio, so the chart below plots them all: higher means faster growth, further left means a calmer ride. On ETH, the dial is remarkable: the Sharpe ratio stays at 0.85 or better across every setting from pure ETH all the way to 60% RiskOFF, while the deepest drawdown shrinks from −79% to −47% and the Calmar ratio peaks mid-dial at 0.72. Across that whole range, the dial changes how much pain you take, not how well you are paid for the risk. On BTC, the same dial charges a small fee per notch—0.81 at zero ballast easing to 0.73 at 60%—mostly because RiskOFF BTC earned 8.2% a year to RiskOFF ETH’s 11.9%. The safety is real on both; ETH happened to hand it out free.

The crash test

Averages hide the nights that matter, so here are the five worst storms of the window, each measured peak to trough inside its own weeks. The pattern repeats in every row: each 60/40 cut its crash by roughly a quarter to a third against its raw asset, because RiskOFF’s floor held its band in all five. The classic 60/40 fell least in the macro events—that is what a mature, low-volatility portfolio does—but its own worst event of the era, 2022, is not on this list, because 2022 was not a crypto crash. It was the year the classic recipe’s ballast failed: while stocks and bonds sank together and the classic 60/40 lost 15.8%, the BTC 60/40 turned BTC’s −64.4% year into −47.6%, and the ETH 60/40 turned ETH’s −66.8% into −44.3%.

Deepest peak-to-trough fall inside each event window. BTC 60/40 = 60% BTC, 40% RiskOFF BTC; ETH 60/40 = 60% ETH, 40% RiskOFF ETH; both rebalanced monthly.

The recovery race

Drawdowns are quoted in percent but lived in time, and time is where the ballast earns its keep most visibly. After the November 2021 peak, ETH spent 1,374 days underwater—its holders were not whole again until August 2025. The ETH 60/40 was whole by March 2024, 17 months earlier. It fell less to begin with, and through every month of the bear its calendar was mechanically selling ballast to buy more ETH at lower prices. On BTC the same race ended in a five-day tie—841 days for the 60/40 against 846 for BTC—which is its own honest lesson: the shallower the crash, the less the ballast has to do.

Against the original

The ladder below puts everything on the single scale defined at the top. Honesty requires two concessions before the punchline. Gold—quietly one of the great assets of this window—beat every raw asset and the BTC 60/40 on the pain-adjusted Calmar score (0.66); only the ETH 60/40 scored higher (0.71), and gold needed less than a third of its volatility to do it. And the classic 60/40’s deepest drawdown, −21.6%, is far shallower than either crypto version’s: an investor who cannot tolerate a −59% fall should not hold a crypto portfolio, with or without ballast. What the comparison does establish is narrower and still remarkable: measured per unit of risk taken, both crypto 60/40s beat the portfolio construct they borrowed the name from—0.87 and 0.78 against 0.56—in the same six years, on the same yardstick, with the same calendar.

What the 40 does not do

• It does not make crypto mild. Falls of −59% and −61% are still crypto-sized. The ballast turns a catastrophic ride into a survivable one; it does not turn BTC into a bond.
• It does not diversify. RiskOFF moves 0.77 to 0.78 correlated with its own asset—it is the same bet with the amplitude turned down, which is precisely what lets it stay useful inside a crypto-only portfolio. It de-risks; it does not diversify.
• It is not free everywhere. The ETH 60/40 kept its asset’s score; the BTC 60/40 gave up three hundredths of Sharpe for its calm. The honest summary is a dial that is free on one asset and fairly priced on the other—not a free lunch everywhere.

The other side

This is the third study in a series. In The Leverage Tax we measured what crypto’s favorite offense actually costs: holding a 2X leveraged BTC position through perpetual futures burns roughly 11.6% a year in funding fees alone, before the market even moves. In Calm Engineered we measured the defense: RiskOFF ran calmer than gold and the S&P 500 over the same six years. This study puts the two sides where they belong—in one portfolio, doing the oldest job in finance: growth on one side, ballast on the other, a calendar in between. As per our backtests, that construction kept nearly all of the era’s best risk-adjusted score on ETH, bought its calm at a fair price on BTC, and surfaced from the bear a year and a half sooner on the asset that needed it most.

Crypto did not need better coins. It needed ballast. The missing 40 finally has a design.

¹ Daily returns for BTC, ETH, RiskOFF BTC, and RiskOFF ETH from The Risk Protocol’s pricing engine (net token value), chain-linked across epochs: 2,382 daily observations from 2019-12-31 through 2026-07-08. ↩

² S&P 500 (SPY) and US bonds (AGG, the Bloomberg US Aggregate ETF) as dividend-adjusted daily closes from Yahoo Finance; gold as the COMEX continuous front-month contract; cash as the 3-month US Treasury constant-maturity yield (FRED series DGS3MO). Crypto trades every day of the year and US markets do not; each series is scored on its own trading days, and re-scoring the crypto portfolios on NYSE days only moves their Sharpe ratios by at most a hundredth (0.87 and 0.77 against the classic’s 0.56). ↩

³ RiskOFF is minted by splitting a BTC or ETH deposit into two tokens: RiskOFF keeps a floor under each epoch’s losses in exchange for a cap on its gains, and the balance of the exposure goes to the other token in the pair. Its behavior over these six years—volatility below gold’s and the S&P 500’s—is measured in Calm Engineered. ↩

⁴ Bloomberg US Aggregate total return of roughly −13% in 2022, the worst calendar year in the index family’s history and, per Charlie Bilello’s long-run series as cited by The Motley Fool, the worst year for US bonds since 1937. ↩

⁵ All portfolio results are backtests on engine net token values; live results would also reflect execution, gas, and slippage. Trading costs barely move the result: charging 25 basis points on every rebalance trade shifts the ETH 60/40’s ending value from $9.77 to $9.70 and the BTC 60/40’s from $5.44 to $5.41. The monthly calendar is a neutral convention, not an optimization: in this window every cadence from weekly to annual beat never rebalancing on both portfolios—letting the engine swell unchecked dragged the deepest falls to −72% (BTC) and −76% (ETH), most of the way back to holding the raw asset. No signals, forecasts, or discretion anywhere: the calendar is the entire strategy. ↩