Similar Sharpe ratios hide severe loss tails and distort position sizing
We compare the loss tails of a put-selling index, the US stock market, the momentum factor and fourteen hedge fund strategy indices with what a normal model predicts. Equal volatility hid up to a fourfold difference in drawdown. Expected shortfall describes the tail better, but estimated from one history it did not beat volatility sizing on average out of sample.
September 2026
BlackRidge Research
BlackRidge Research: Asymmetric Tail Risk01
BlackRidge
Tail Risk Sizing
02 / 13
Put-sellers and the market earned similar Sharpe ratios, yet the put-seller's bad months fell 1.50 times deeper than normal models predict
The put-selling index comes within 0.09 of the stock market's Sharpe ratio [02]. The symmetry fails in the tail: during the worst 5% of months, the strategy's average loss hits 1.50 times the normal model's prediction (against 1.18 for the market). Because most hedge fund indices share this negative skew, volatility and Sharpe ratios are the wrong inputs for sizing leveraged positions [04]. Size on the loss tail and a drawdown limit, treating the tail estimate as uncertain.
Sharpe ratio comparison
0.56 / 0.65
The put-selling index against the US stock market.
Tail loss multiplier
1.50
Realised CVaR versus normal expected shortfall for put sellers.
Negative skewness
10 / 14
Hedge fund strategies exhibiting an asymmetric downside.
Drawdown variance
15–60%
Observed across strategies levered to the same volatility target.
Put-selling index
1.50
US stock market
1.18
Momentum factor
1.37
Hedge fund average
1.24
Realised CVaR divided by the normal model's expected shortfall (1.00 means the normal model is exactly right). Uses monthly data from 2007 to 2026 for the first three assets, and 1997 to 2026 for the hedge fund average [01][02][04].
Evidence
Fourteen hedge fund indices scaled to a 10% annual volatility produce largest drawdowns ranging from -15.3% to -59.8% [04]. Identical volatility numbers mask entirely different tail risks.
Interpretation
Many hedge fund strategies generate payoffs resembling a short put on the market [09]. Mean-variance analysis and volatility metrics ignore this structural left-tail risk entirely.
Decision implication
Volatility-based sizing is incomplete for negatively skewed strategies: add a drawdown limit tested on stressed history and use CVaR as a cross-check.
Caveat
The CVaR figure rests on just 11 worst months and is inherently noisy. Out of sample, CVaR sizing did not beat volatility sizing on average.
BlackRidge Research: Asymmetric Tail Risk02
BlackRidge
Gaussian Yardsticks
03 / 13
105 days fell more than four standard deviations. A normal model expects one.
Parametric 95% VaR stops at 1.645 standard deviations. It says absolutely nothing about how far beyond that mark the losses go. In the middle of the distribution the normal model is too pessimistic, yet in the far tail it becomes absurdly optimistic [01].
Number of days (log scale)
Observed daysNormal model
Count of daily market losses exceeding standard deviation thresholds, 1 July 1926 to 31 August 2026, US market data [01]. Loss threshold in standard deviations
Four-sigma loss events
105 / 0.8
Against 0.83 expected under a normal model.
October 1987 crash severity
16.2σ
Measured in daily standard deviations.
Failures of rolling 99% VaR
2.36%
Breach rate for a 1% target model.
Evidence
The US market registered a loss beyond five standard deviations on 51 separate days [01]. A normal model expects 0.0075 days.
Interpretation
Value-at-Risk is not a coherent risk measure because it can penalise diversification [07]. Expected shortfall is a coherent measure.
Caveat
These numbers rely on daily index data before transaction costs. The 1926 to 1940 period weighs heavily in the far tail.
BlackRidge Research: Asymmetric Tail Risk03
BlackRidge
TAIL RISK COMPARISON
04 / 13
The put-seller came within 0.09 of the market's Sharpe ratio and lost 5.9 standard deviations in one month.
The index collects option premium month after month and gives it back in a few short months [02]. Its volatility runs at two thirds of the US market's [01]. That lower variance flatters the Sharpe ratio.
Value of $1
Cboe S&P 500 PutWrite IndexUS Stock Market
Growth of 1 dollar, log scale, monthly, February 2007 to August 2026 [01][02].
Measure
Put-selling index
US stock market
Return a year
7.1%
11.1%
Volatility a year
10.7%
15.9%
Sharpe ratio
0.56
0.65
Skewness
−1.67
−0.51
Excess kurtosis
7.1
1.0
Normal 95% VaR, a month
−4.47%
−6.55%
Realised CVaR, a month
−8.66%
−10.03%
Normal expected shortfall, a month
−5.77%
−8.47%
Worst month
−17.7%
−17.1%
Largest drawdown
−32.7%
−50.3%
Months under water
30 mo
52 mo
Interpretation
The normal model reads a calm track record as low risk. It completely misses that the put-seller's worst month was a 5.9 standard deviation move.
Caveat
This comparison covers one crisis-heavy period, and the stock market had the deeper drawdown in percent [01][02]. The focus here is the asymmetric shape of the risk, not the absolute size of the loss.
BlackRidge Research: Asymmetric Tail Risk04
BlackRidge
Hedge funds: negative skew is the rule
05 / 13
Ten of fourteen hedge fund strategies have a negative skew, and their tails run up to 1.53 times the normal forecast
Arbitrage strategies take small steady gains and pay in rare large losses. Their Sharpe ratios often look best, but for the arbitrage strategies part of the steady return is payment for carrying the left tail [09]. The more negative the skew, the further the realised tail exceeds the normal forecast, driving a correlation of -0.85 across indices [04].
Indices with negative skewness
10 / 14
Out of the fourteen BarclayHedge strategies tracked.
Correlation with tail ratio
−0.85
How closely negative skew tracks the failure of normal risk forecasts.
Fixed income arbitrage tail event
11.3σ
Size of the loss during October 2008 in standard deviations.
Strategy
Skewness
Excess kurtosis
Normal expected shortfall
Realised CVaR
Ratio
Fixed income arbitrage
−4.85
47.1
−2.10%
−3.22%
1.53
Convertible arbitrage
−2.65
23.5
−2.72%
−3.43%
1.26
Multi-strategy
−1.90
10.1
−2.14%
−3.03%
1.42
Merger arbitrage
−1.42
12.0
−1.69%
−2.19%
1.30
Event driven
−1.38
7.5
−3.51%
−4.46%
1.27
Distressed securities
−1.18
3.7
−3.39%
−4.75%
1.40
Fund of funds
−0.80
4.6
−2.74%
−3.54%
1.29
All strategies (composite)
−0.73
3.6
−3.50%
−4.46%
1.27
Emerging markets
−0.72
3.9
−6.97%
−8.50%
1.22
Equity long bias
−0.61
1.8
−5.87%
−7.07%
1.20
Equity market neutral
0.01
1.8
−1.26%
−1.45%
1.15
Technology sector
0.53
3.1
−6.80%
−7.02%
1.03
Global macro
0.57
1.2
−2.73%
−2.50%
0.92
Equity long/short
0.67
4.6
−3.25%
−3.41%
1.05
Evidence
The realised CVaR divided by the normal expected shortfall averages 1.24 across the fourteen hedge fund indices, hitting 1.53 for fixed income arbitrage [04]. Global macro is the only strategy where this ratio sits below one [04].
Decision implication
Discard normal VaR when sizing arbitrage allocations. Allocators must scale positions to survive historical drawdowns and realised CVaR instead of relying on Sharpe ratios.
Caveat
Self-reported returns from surviving funds and smoothed illiquid pricing understate both volatility and the loss tail [04]. Re-ranking the indices by CVaR changes little at the top, moving just one strategy three places [04].
BlackRidge Research: Asymmetric Tail Risk05
BlackRidge
DURATION OF LOSS
06 / 13
Momentum has been under its 2008 peak for 213 months. The put-seller needed 30.
Volatility and the Sharpe ratio have no time dimension. Months spent below the high-water mark decide whether investors stay. That duration dictates whether a leveraged book survives margin calls [01][02].
Drawdown from previous peak, %, monthly, February 2007 to August 2026 [02][01].
Put-selling recovery
30 mo
Months under water from 2008 to 2010.
US market recovery
52 mo
Months spent below the October 2007 peak.
Momentum underwater
213 mo
Months failing to reach the November 2008 peak.
Evidence
The put-writing index cleared its May 2008 high within 30 months [02]. The momentum factor broke down in November 2008 and has spent 213 months failing to recover [01].
Interpretation
Short-volatility drops violently but rebuilds capital through constant premium collection. Momentum crashed in the 2009 panic rebound and stayed down.
Decision implication
Size positions against your tolerance for time under water. A shallow expected shortfall matters little if recovery takes two decades.
Caveat
Drawdown length depends heavily on the start date. A single crisis skews the historical recovery timeline.
BlackRidge Research: Asymmetric Tail Risk06
BlackRidge
MOMENTUM CRASHES
07 / 13
Momentum showed a Sharpe ratio of 0.85 just before losing 49% in three months
The trailing record looked perfectly calm, with volatility at an ordinary 14.2% in February 2009 [01]. The crash came when the market rebounded after the fall. Daniel and Moskowitz document this pattern [10].
Drawdown (%)
Momentum factor drawdown from previous peak, %, monthly, January 1927 to August 2026 [01].
Trailing 36-month Sharpe ratio
0.85
Measured up to February 2009.
Three-month loss
−49.4%
Realised from March to May 2009.
Months under water
293 mo
From June 1932 to December 1956.
Interpretation
The high trailing Sharpe ratio of February 2009 carried no warning about the crash that followed. A smooth recent history simply cannot forecast the left tail.
Decision implication
Discard trailing volatility when sizing momentum allocations. Anchor your exposure to the severe drawdown you will face when a fallen market sharply rebounds.
Caveat
Momentum is a long-short academic factor before costs, not an investable fund. A trader can partly anticipate the danger since its crashes cluster in market rebounds [10].
BlackRidge Research: Asymmetric Tail Risk07
BlackRidge
ASYMMETRIC RECOVERY
08 / 13
At three times leverage the put-seller fell 76% and needed a 321% gain to recover
Leverage scales the annual return roughly linearly. The gain needed to get back grows much faster than the initial loss. At 1x leverage the index needed a 48% gain to recover and at 3x it needed 321%, while the months spent under water rose only from 30 to 46 [02].
The arithmetic of compounding punishes deep drawdowns severely. Time under water increases modestly, but the performance burden required to repair the capital base explodes.
Decision implication
Size positions based on the gain required to survive a tail event, not just expected volatility. A theoretical recovery path means nothing if the intervening loss breaches the risk tolerance.
Caveat
These calculations assume financing at the Treasury bill rate without margin calls. In practice a 58% loss at 2x leverage would trigger a forced sale near the bottom, locking the loss in so the recovery shown in the table would not happen.
BlackRidge Research: Asymmetric Tail Risk08
BlackRidge
What volatility sizing hides
09 / 13
At the same 10% volatility, drawdowns ran from 15% to 60%
Equal volatility budgets gave fixed income arbitrage four times the drawdown of global macro. A 20% drawdown limit would have allowed fixed income arbitrage just 0.30 of its volatility-based leverage. Global macro could have taken 1.30. Estimated on 1997-2011 and applied to 2011-2026, CVaR sizing lowered the drawdown for 10 of 14 indices but left the average almost unchanged [04].
fixed income arbitrage
−59.8%
convertible arbitrage
−51.8%
distressed securities
−48.9%
fund of funds
−41.1%
multi-strategy
−40.1%
all strategies composite
−33.6%
equity long/short
−21.8%
global macro
−15.3%
Largest drawdown, %, each index levered to 10% annual volatility over its full history, monthly, January 1997 to June 2026 [04].
Fixed income leverage
0.30
Share of volatility leverage under a 20% drawdown limit.
Improved by CVaR
10 / 14
Indices with a smaller out-of-sample drawdown under CVaR sizing.
Average future drawdown
−17.1% / −17.2%
Comparing CVaR against volatility sizing out of sample.
Evidence
Set hedge fund indices to a uniform 10% volatility, and outcomes scatter wildly. Fixed income arbitrage suffered a 59.8% drawdown. Global macro contained losses to 15.3%.
Interpretation
A simple volatility target misses the shape of the loss tail. Variance treats all deviations equally, blinding the model to strategies that hide risk in rare events.
Decision implication
Size on a drawdown limit tested on stressed history, use CVaR as a cross-check, and haircut both [07][08]. Relying on a single metric invites failure.
Caveat
The full-sample result uses hindsight. The out-of-sample test covers one data split and a largely calm decade.
BlackRidge Research: Asymmetric Tail Risk09
BlackRidge
ESTIMATION NOISE
10 / 13
The put-seller's CVaR estimate swung from -9.2% to -6.1% when October 2008 left the window
CVaR describes the left tail better than VaR, yet it demands a lot of data. A ten-year rolling window leaves you looking at a handful of extreme months [11]. The realised tail proved worse than the normal forecast in all 116 windows we checked, and dropping the 2008 crash shifted the severity estimate by a third [02].
Monthly return (%)
Realised CVaRNormal expected shortfall
rolling ten-year 95% CVaR and normal expected shortfall, % a month, windows ending January 2017 to August 2026, [02]
90% CVaR interval
−11.7% … −6.0%
Bootstrap bounds on the estimate
90% Sharpe ratio interval
0.16 … 1.01
Range of possible risk profiles
Tail observations
5
Months driving the 97.5% metric
Interpretation
A single history gives a wide interval, so a CVaR-based size carries that uncertainty. Tail events barely appear in this sample.
Decision implication
A 97.5% expected shortfall computed on these 235 monthly returns rests on 5 observations. This is why a single fund record needs stress scenarios and pooled histories [05].
Caveat
The bootstrap assumes past blocks remain a valid guide. If the market regime breaks, the true tail will land entirely outside these resampled bounds.
BlackRidge Research: Asymmetric Tail Risk10
BlackRidge
ASYMMETRIC TAIL RISK
11 / 13
The put-seller's three-year Sharpe ratio is higher than in 87.5% of months since 2010 while VIX sits below its median
Calm markets routinely pay insurance sellers, with VIX closing above the subsequent realised volatility on 85.2% of days [03]. This volatility premium makes their record look best exactly when the stored-up tail is least visible [02]. We saw the same setup when VIX traded at 16.23 at the end of June 2007.
Trailing 36-month Sharpe ratio
Cboe S&P 500 PutWrite IndexUS stock market
Trailing 36-month Sharpe ratio, monthly, January 2010 to August 2026 [02][01].
VIX on 25 September 2026
14.87
Lower than 69% of trading days since January 1990.
Historical volatility premium
85.2%
Trading days since 1990 VIX exceeded subsequent realised volatility.
EU market risk rules start
2027
Transition to capital requirements based on expected shortfall.
Interpretation
High Sharpe ratios on short-volatility strategies mostly reflect a quiet market. Investors mistake the steady premium collection for alpha right up until the loss tail materialises.
Decision implication
Global banking regulators are shifting market risk measures to 97.5% expected shortfall [05][06]. Allocators judging leveraged strategies should demand this same tail metric and the full drawdown record instead of stopping at the Sharpe ratio.
BlackRidge Research: Asymmetric Tail Risk11
BlackRidge
Verdict
12 / 13
Standard risk models severely understate the loss tail in leveraged and short-volatility strategies.
Standard risk metrics routinely disguise the structural downside of short-volatility payoffs. Prudent capital allocation demands assessing historical drawdowns instead of trusting a normal model built around a 1.08% daily standard deviation.
01
105 days exceeded four standard deviations.
02
Selling puts yielded comparable risk metrics alongside a realised expected shortfall 1.50 times worse than normal estimates.
03
Leverage accelerates returns but heavily compounds the downside, leaving a levered put-writing strategy needing a 321% gain just to recover from its trough.
04
Momentum strategies look attractive during stable markets. Then they collapse and trap capital for an entire generation. The factor is still under its historical peak after 213 months [01].
05
Scaling hedge fund indices to identical volatility produced maximum drawdowns ranging from 15% to 60% [04]. Relying on historical CVaR for position sizing merely matched simple volatility weighting out of sample.
Tail Risk Summary
Measure
What it shows
Value
Extreme tail days
Count of daily losses far outside normal distribution limits
105
Put-writing shortfall ratio
Ratio of realised extreme losses against normal model predictions
1.50
Negative skewness strategies
Count of hedge fund indices exhibiting asymmetric downside risk
10 / 14
Momentum drawdown duration
Months spent failing to recover the previous market peak
213 mo
Levered recovery requirement
Return required to break even following the maximum trough
+321%
Volatility targeted drawdowns
Range of maximum losses across strategies scaled to identical volatility
15–60%
Never allocate capital to a leveraged strategy without examining its loss tail and the duration of its worst drawdown. Demand to see exactly how the proposed position size would have behaved through that entire historical trough.
Expected shortfall requires much larger data samples than VaR to achieve the same estimation accuracy.
This report is research rather than investment advice. Index returns stand gross of fees, costs and taxes, and the hedge fund indices run entirely on self-reported data. Past market tails do not bound future ones.
BlackRidge Research: Asymmetric Tail Risk13
Citation context
The put-selling index comes within 0.09 of the stock market's Sharpe ratio. In the worst 5% of months, its average loss is 1.50 times the normal model's prediction, against 1.18 for the market. At the same 10% volatility, drawdowns ranged from 15% to 60%.
Sample and method: loss tails of a put-selling index, the US stock market, the momentum factor and fourteen hedge fund strategy indices, 1926 to 2026.
Limits: the CVaR figure rests on just 11 worst months and is inherently noisy. Out of sample, CVaR sizing did not beat volatility sizing on average.
When you use a result that another publication established, cite that original work; it is linked in Sources. Cite this report for our synthesis, explanation or an identified recalculation. No link is required in return.
BlackRidge is an independent research bureau introducing private investors to
quantitative traders through multiple strategy providers. We publish quantitative
research. Investors access strategies through PAMM accounts at the broker. We are
not a fund or broker and never hold client money.
01
Your funds stay at the broker. The account is opened in your own name. BlackRidge never receives, holds, or has withdrawal rights over client capital.
02
No fee on your profits. A one-time access payment of 6.7% of the agreed trading level, paid to the selected strategy provider. No management fee, no performance fee, no profit share, in any year.
03
You fund the risk, not the exposure. Allocations are notionally funded: you agree a trading level and fund the margin and drawdown allowance behind it. Trading losses can exceed the deposit without applicable negative balance protection; the separate 6.7% access payment is nonrefundable.
We use AI models to gather and aggregate source material and to help prepare each
report. Read it with its cited sources, sample, methods and limitations, and send
corrections through research methodology and corrections.
The client pays the selected strategy provider a one-time access payment of 6.7% of
the agreed trading level. BlackRidge receives an introduction fee from that provider
and does not receive broker compensation. About BlackRidge.
Minimum funded capital $25,000. The strategy, its live records, the broker,
the selected strategy provider and the full commercial terms are presented on an
introductory call and confirmed in writing before any payment or deposit.
This report is published for information only. It is not investment advice and does
not take account of your circumstances. Trading leveraged instruments including CFDs
carries substantial risk and is not suitable for all investors; you may lose the
capital you fund and, depending on your broker's terms, may owe more than your
deposit. Findings in this report may combine third-party evidence, historical
calculations and illustrative scenarios; they are not verified live trading results
of any strategy provider. Past performance does not indicate future results.