inthesingularity.ai
Finance · Portfolio backtest

F.001What beat holding QQQ, 2011–2026?

Over 15 years, holding only QQQ’s 10 largest stocks compounded at 23.8% a year against 20.3% for QQQ, but fell 40.8% at its worst, leaving its Sharpe ratio level with QQQ’s. Putting a quarter in gold kept most of QQQ’s return, cut the worst drawdown to 27.0% and had the best Sharpe ratio of the five. Bonds, which carried the Permanent and Risk Parity portfolios for a decade, stopped hedging in 2022.

Month-ends Sep 2011 → Sep 2026 · 180 monthly returns · total returns, dividends reinvested
Permanent
25% QQQ · 25% GLD · 25% TLT · 25% T-bills
7.4%
annualized return
Sharpe
0.77
Max DD
−19.1%
Vol
7.8%
Risk Parity
35% QQQ · 25% TLT · 25% IEF · 7.5% GLD · 7.5% DBC
8.6%
annualized return
Sharpe
0.82
Max DD
−22.8%
Vol
8.7%
Equity
100% QQQ
20.3%
annualized return
Sharpe
1.07
Max DD
−32.6%
Vol
17.5%
Equity Gold
75% QQQ · 25% GLD
17.2%
annualized return
Sharpe
1.10
Max DD
−27.0%
Vol
14.1%
QQQ Top 10
QQQ’s top 10 holdings, reset each Sept 30
23.8%
annualized return
Sharpe
1.08
Max DD
−40.8%
Vol
20.5%

Growth of $100

Month-end values for the selected window, log scale. Hover or drag across the chart for values.

PermanentRisk ParityEquityEquity GoldQQQ Top 10

Drawdown from the prior peak

Month-end values within the selected window. Intra-month lows went deeper; see the daily comparison below.

PermanentRisk ParityEquityEquity GoldQQQ Top 10

Risk and return scorecard

For the selected window. The best value in each comparable row is underlined. Sharpe and Sortino use monthly returns over T-bills (the BIL ETF), annualized.

MetricPermanentRisk ParityEquityEquity GoldQQQ Top 10
Annualized return7.4%8.6%20.3%17.2%23.8%
Cumulative return193.0%242.6%1504.3%986.7%2353.1%
Volatility (ann.)7.8%8.7%17.5%14.1%20.5%
Sharpe ratio0.770.821.071.101.08
Sharpe ±1 std. error±0.26±0.26±0.26±0.26±0.26
Sortino ratio1.271.331.871.981.92
Max drawdown−19.1%−22.8%−32.6%−27.0%−40.8%
Max DD peak → troughDec 2021 → Oct 2022Dec 2021 → Oct 2022Dec 2021 → Dec 2022Dec 2021 → Sep 2022Nov 2021 → Dec 2022
Recovered (months underwater)Mar 2024 (27 mo)Aug 2024 (32 mo)Dec 2023 (24 mo)Jul 2023 (19 mo)Nov 2023 (24 mo)
Calmar (return ÷ max DD)0.390.380.620.640.58
Worst month−6.3%−7.9%−13.6%−10.7%−16.2%
Worst month dateApr 2022Apr 2022Apr 2022Apr 2022Apr 2022
Worst rolling 12 months−17.8%−22.1%−32.6%−25.3%−40.4%
Positive months64%61%64%67%65%
Beta to QQQ0.290.411.000.771.10
Correlation to QQQ0.660.821.000.960.94
Drawdown at Sep 30, 2026−3.4%−2.5%0.0%−1.2%−1.5%

The standard-error row (Lo, 2002) shows how loosely each Sharpe ratio is pinned down. One-year volatility and Sharpe rest on twelve monthly observations.

Concentration paid, and cost more risk

The QQQ Top 10 buys QQQ’s ten largest holdings every September 30, at their QQQ weights scaled up to 100%, and holds them for a year. It beat QQQ in 12 of 15 years and added +3.5% a year over 15 years. The extra return came with extra risk: a beta of 1.10 to QQQ, bigger moves both ways (115% of QQQ’s average up month, 112% of its down month), and a 40.8% drawdown in 2022 against 32.6% for QQQ. Its Sharpe ratio, 1.08, is level with QQQ’s 1.07.

Gold was the hedge that held

Gold’s correlation with QQQ stayed near zero throughout: +0.13 from 2011 to 2021 and +0.09 over the last five years. A 25% gold sleeve cut volatility from 17.5% to 14.1% and the worst drawdown from 32.6% to 27.0%, and Equity Gold got back to its 2021 peak in 19 months against 24 for QQQ.

Bonds stopped hedging in 2022

From 2011 to 2021, long Treasuries tended to rise when QQQ fell: their monthly correlation was −0.23. Over the last five years it was +0.54. Risk Parity, with half its money in Treasuries, went from a Sharpe ratio of 1.36 in its first ten years to 0.24 in the last five and took 32 months to recover from 2022. The Permanent Portfolio fell from 1.02 to 0.46; its T-bills and gold softened the blow.

The last year went the other way

The September 2025 list returned +12.2% over the following twelve months against +23.8% for QQQ. Netflix fell 42.0%, Tesla 20.2% and Microsoft was flat, while the semiconductor rally ran through Micron and AMD, which were outside the top 10. Both are in the current list.

QQQ Top 10 against QQQ

Each September 30 the portfolio buys QQQ’s 10 largest holdings from that year’s annual report, scales their combined QQQ weight to 100%, and holds them for twelve months while the weights drift with prices. Alphabet’s two share classes count as two holdings, as they do in QQQ’s list.

12 of 15years ahead of QQQ
22different stocks held
52.5%of QQQ in its top 10, on average
8.5%average yearly turnover
WindowReturn vs QQQ, a yearTracking errorInformation ratioUp-month captureDown-month captureMonths ahead
15Y+3.5%7.2%0.48115%112%61%
5Y+4.4%10.3%0.46120%113%60%
3Y+6.0%11.6%0.47119%116%64%
1Y−11.6%16.2%−0.5982%114%42%

Capture is the top 10’s average return in QQQ’s up (or down) months as a share of QQQ’s. Above 100% on both sides means bigger moves in both directions.

Year by year

The list taken each September 30 and how it did over the next twelve months. Weights are percent of the top-10 portfolio.

List dateTop 10QQQDifferenceTurnoverChangesHoldings and weights
Sep 2011+40.2%+31.9%+8.3%—StartAAPL 27.0 · MSFT 15.9 · ORCL 11.1 · GOOGL 10.0 · INTC 8.6 · AMZN 7.5 · CSCO 6.5 · QCOM 6.2 · AMGN 3.9 · CMCSA 3.3
Sep 2012+1.0%+16.7%−15.7%5.4%+KFT −AMGNAAPL 34.5 · MSFT 13.8 · GOOGL 10.9 · ORCL 8.5 · AMZN 6.3 · INTC 6.3 · QCOM 5.9 · CSCO 5.6 · CMCSA 4.2 · KFT 4.0
Sep 2013+38.3%+27.1%+11.2%13.7%+GILD, META −KFT, ORCLAAPL 25.0 · MSFT 16.0 · GOOGL 13.9 · AMZN 8.2 · CSCO 7.2 · QCOM 6.7 · INTC 6.6 · CMCSA 5.6 · GILD 5.6 · META 5.3
Sep 2014+6.3%+4.1%+2.2%11.4%+GOOG −CMCSAAAPL 26.9 · MSFT 17.0 · GOOG 8.7 · INTC 7.7 · GOOGL 7.4 · GILD 7.2 · META 7.1 · AMZN 6.6 · CSCO 5.7 · QCOM 5.6
Sep 2015+24.0%+18.0%+6.0%6.8%+CMCSA −QCOMAAPL 26.6 · MSFT 15.0 · AMZN 10.1 · GOOG 8.9 · META 8.6 · GOOGL 7.8 · GILD 6.1 · INTC 6.1 · CSCO 5.7 · CMCSA 5.1
Sep 2016+25.1%+23.7%+1.3%6.3%+AMGN −GILDAAPL 21.2 · MSFT 15.6 · AMZN 13.8 · META 10.3 · GOOG 9.3 · GOOGL 8.2 · INTC 6.2 · CSCO 5.5 · CMCSA 5.5 · AMGN 4.3
Sep 2017+42.2%+28.7%+13.5%1.5%No changeAAPL 22.6 · MSFT 16.3 · AMZN 13.1 · META 11.5 · GOOG 9.5 · GOOGL 8.2 · CMCSA 5.1 · INTC 5.1 · CSCO 4.8 · AMGN 3.9
Sep 2018+1.1%+2.5%−1.4%8.3%+NFLX, NVDA −AMGN, CMCSAAAPL 22.3 · AMZN 19.9 · MSFT 17.9 · GOOG 8.5 · META 8.1 · GOOGL 7.4 · CSCO 4.7 · INTC 4.5 · NVDA 3.5 · NFLX 3.3
Sep 2019+53.3%+48.3%+5.0%8.7%+CMCSA, PEP −NFLX, NVDAMSFT 20.8 · AAPL 19.9 · AMZN 16.9 · META 8.5 · GOOG 8.4 · GOOGL 7.4 · INTC 5.0 · CSCO 4.6 · CMCSA 4.5 · PEP 4.2
Sep 2020+34.6%+29.5%+5.1%18.7%+ADBE, NVDA, PYPL, TSLA −CMCSA, CSCO, INTC, PEPAAPL 23.8 · MSFT 19.2 · AMZN 19.0 · META 7.6 · TSLA 6.1 · GOOGL 6.1 · GOOG 5.9 · NVDA 5.1 · ADBE 3.6 · PYPL 3.6
Sep 2021−24.0%−24.9%+0.8%2.8%No changeAAPL 20.8 · MSFT 18.9 · AMZN 14.8 · TSLA 8.6 · GOOG 7.6 · META 7.2 · GOOGL 7.2 · NVDA 7.0 · PYPL 4.1 · ADBE 3.7
Sep 2022+39.9%+35.0%+4.9%9.8%+COST, PEP −ADBE, PYPLAAPL 25.5 · MSFT 19.9 · AMZN 13.2 · TSLA 9.5 · GOOG 6.8 · GOOGL 6.6 · META 5.3 · NVDA 4.9 · PEP 4.3 · COST 4.0
Sep 2023+57.3%+37.2%+20.1%8.0%+AVGO −PEPAAPL 22.4 · MSFT 19.6 · AMZN 11.0 · NVDA 9.0 · META 7.8 · TSLA 6.6 · GOOGL 6.5 · GOOG 6.4 · AVGO 6.2 · COST 4.5
Sep 2024+36.6%+23.7%+13.0%5.0%No changeAAPL 17.8 · MSFT 16.1 · NVDA 15.0 · AVGO 10.4 · META 10.0 · AMZN 9.8 · TSLA 6.4 · COST 5.1 · GOOGL 4.9 · GOOG 4.7
Sep 2025+12.2%+23.8%−11.6%12.1%+NFLX −COSTNVDA 18.7 · MSFT 15.9 · AAPL 15.6 · AVGO 10.6 · AMZN 9.6 · TSLA 6.7 · META 6.6 · GOOGL 5.8 · GOOG 5.4 · NFLX 5.2

How the September 2025 list did

Return from September 30, 2025 to September 30, 2026, and each stock’s contribution to the portfolio’s +12.2%.

StockWeight at start12-month returnContribution
NVDA18.7%+22.6%+4.2%
MSFT15.9%−0.2%0.0%
AAPL15.6%+31.3%+4.9%
AVGO10.6%+7.2%+0.8%
AMZN9.6%+13.5%+1.3%
TSLA6.7%−20.2%−1.3%
META6.6%−0.9%−0.1%
GOOGL5.8%+41.9%+2.4%
GOOG5.4%+40.2%+2.2%
NFLX5.2%−42.0%−2.2%

What it holds next

QQQ’s top 10 on October 16, 2026, scaled to 100%: the stand-in for the September 30, 2026 list, which QQQ publishes later in the year. Micron and AMD come in; Broadcom and Netflix drop out. The last three rows are next in line.

StockQQQ weightPortfolio weightOct 1–16
NVDANVIDIA8.14%17.3%−6.2%
AAPLApple7.91%16.9%+3.1%
MSFTMicrosoft6.20%13.2%+3.1%
MUMicron Technology4.32%9.2%−17.5%
AMZNAmazon4.25%9.1%+0.3%
AMDAdvanced Micro Devices3.76%8.0%−13.6%
METAMeta Platforms3.33%7.1%−2.2%
GOOGLAlphabet Class A3.15%6.7%−1.0%
TSLATesla2.94%6.3%0.0%
GOOGAlphabet Class C2.94%6.3%−1.1%
SPCXSpaceX2.73%——
WMTWalmart2.52%——
AVGOBroadcom2.47%——

Where the return came from

Each stock’s dollar gain per $100 invested in September 2011, summed over the years it was in the list. The rows add up to the portfolio’s total gain.

StockYears in listSpanGain per $100Share of gain
AAPL152011–2026$49120.9%
NVDA72018–2026$46219.6%
MSFT152011–2026$33914.4%
AVGO32023–2026$24110.2%
AMZN152011–2026$2098.9%
META132013–2026$1827.7%
GOOGL152011–2026$1797.6%
GOOG122014–2026$1616.8%
TSLA62020–2026$783.3%
COST32022–2025$381.6%
INTC92011–2020$170.7%
CSCO92011–2020$150.6%
CMCSA72011–2020$100.4%
AMGN32011–2018$60.2%
PEP22019–2023$30.1%
GILD32013–2016$20.1%
ORCL22011–2013$20.1%
QCOM42011–2015$10.0%
KFT12012–2013$10.0%
ADBE22020–2022−$14−0.6%
PYPL22020–2022−$18−0.8%
NFLX22018–2026−$52−2.2%

Stress episodes

Return over each episode, month-end to month-end. Shading scales with the size of the move.

EpisodePermanentRisk ParityEquityEquity GoldQQQ Top 10
Taper tantrum + gold crash, Apr–Jun 2013−6.2%−3.5%+3.7%−3.0%+3.0%
Q4 2018 selloff, Oct–Dec 2018−1.0%−4.6%−16.7%−10.7%−20.6%
COVID crash, Feb–Mar 2020+0.1%−0.9%−12.9%−9.9%−12.9%
2022 inflation & rate shock, Jan–Dec 2022−16.6%−22.1%−32.6%−25.3%−40.4%
2023 long-bond rout, Aug–Oct 2023−5.5%−8.2%−8.4%−6.1%−7.5%
2025 tariff shock, Feb–Mar 2025+1.7%−0.6%−10.1%−4.5%−16.0%
March 2026−5.3%−3.0%−4.8%−6.7%−4.4%
June–July 2026−4.8%−4.1%−6.7%−7.1%−5.9%
October 2026 so far (Oct 1–16)0.0%+0.2%−4.0%−3.0%−3.6%

What each ETF did in the same episodes

EpisodeQQQGLDTLTIEFDBC
Apr–Jun 2013+3.7%−22.9%−5.6%−4.1%−8.0%
Q4 2018−16.7%+7.5%+4.6%+3.9%−18.3%
Feb–Mar 2020−12.9%−0.9%+13.4%+6.8%−22.9%
2022−32.6%−0.8%−31.2%−15.2%+19.4%
Aug–Oct 2023−8.4%+1.0%−15.7%−5.7%−0.6%
Feb–Mar 2025−10.1%+11.4%+4.4%+3.2%+2.4%
Mar 2026−4.8%−11.1%−4.2%−2.3%+15.3%
Jun–Jul 2026−6.7%−10.9%−3.4%−1.1%−0.1%
Oct 1–16, 2026−4.0%+0.3%+3.3%+1.6%+4.4%

Worst drawdowns over 15 years

The three deepest peak-to-trough declines for each portfolio on month-end values, and the months spent below the prior peak.

Permanent

DepthPeakTroughRecoveredMonths
−19.1%Dec 2021Oct 2022Mar 202427
−9.2%Sep 2012Jun 2013Jun 201421
−6.7%Jul 2016Dec 2016May 201710

Risk Parity

DepthPeakTroughRecoveredMonths
−22.8%Dec 2021Oct 2022Aug 202432
−5.6%Sep 2016Nov 2016Apr 20177
−5.4%Aug 2018Dec 2018Mar 20197

Equity

DepthPeakTroughRecoveredMonths
−32.6%Dec 2021Dec 2022Dec 202324
−17.0%Aug 2018Dec 2018Apr 20198
−12.9%Jan 2020Mar 2020Apr 20203

Equity Gold

DepthPeakTroughRecoveredMonths
−27.0%Dec 2021Sep 2022Jul 202319
−11.0%Aug 2018Dec 2018Mar 20197
−9.9%Jan 2020Mar 2020Apr 20203

QQQ Top 10

DepthPeakTroughRecoveredMonths
−40.8%Nov 2021Dec 2022Nov 202324
−21.3%Aug 2018Dec 2018Apr 20198
−16.0%Jan 2025Mar 2025Jun 20255

Month-end data understates fast crashes

QQQ’s deepest declines on daily closes against the month-end figures used here. Every portfolio’s intra-month low was deeper than shown, most for the QQQ-heavy ones.

QQQ drawdownDaily closesMonth-endDifference
2021–22 bear market−35.12%−32.6%−2.5%
COVID crash, 2020−28.56%−12.9%−15.7%
Q4 2018 selloff−22.80%−17.0%−5.8%
2025 tariff shock−22.77%−10.1%−12.7%

Calendar-year returns

Total returns by year. ETF columns appear on wider screens; QQQ alone is the Equity column.

YearPermanentRisk ParityEquityEquity GoldQQQ Top 10GLDTLTIEFDBCBIL
2011 (Oct–Dec)+1.1%+3.0%+6.7%+4.0%+6.7%−3.8%+1.5%+1.3%+4.2%0.0%
2012+7.5%+9.5%+18.1%+15.3%+20.1%+6.6%+2.6%+3.7%+3.5%0.0%
2013−2.9%+3.5%+36.6%+17.8%+29.0%−28.3%−13.4%−6.1%−7.7%−0.1%
2014+10.7%+13.0%+19.2%+13.8%+19.9%−2.2%+27.3%+9.1%−28.1%−0.1%
2015−0.7%+0.4%+9.4%+4.2%+18.5%−10.7%−1.8%+1.5%−27.6%−0.1%
2016+4.7%+5.5%+7.1%+8.0%+8.5%+8.0%+1.2%+1.0%+18.5%+0.1%
2017+13.5%+15.3%+32.7%+27.5%+41.2%+12.8%+9.2%+2.6%+4.9%+0.7%
2018+0.3%−0.4%−0.1%+0.4%+1.4%−1.9%−1.6%+1.0%−11.6%+1.7%
2019+18.1%+21.3%+39.0%+33.8%+44.9%+17.9%+14.1%+8.0%+11.8%+2.0%
2020+23.7%+26.9%+48.6%+43.0%+51.5%+24.8%+18.2%+10.0%−7.9%+0.4%
2021+4.3%+9.7%+27.4%+18.9%+37.9%−4.1%−4.6%−3.3%+41.4%−0.1%
2022−16.6%−22.1%−32.6%−25.3%−40.4%−0.8%−31.2%−15.2%+19.4%+1.4%
2023+18.0%+20.0%+54.9%+43.6%+78.3%+12.7%+2.8%+3.6%−6.2%+4.9%
2024+11.9%+8.5%+25.6%+26.0%+54.5%+26.7%−8.1%−0.6%+2.2%+5.2%
2025+22.2%+15.7%+20.8%+31.5%+23.9%+63.7%+4.3%+8.0%+8.1%+4.2%
2026 (Jan–Sep)+3.6%+7.7%+20.8%+16.0%+9.3%−3.9%−8.0%−4.6%+44.5%+2.7%

The stock–bond hedge flipped

Rolling 36-month correlation of monthly returns with QQQ. Below zero, the asset tended to rise when QQQ fell.

QQQ–TLTQQQ–GLDQQQ–DBC
Monthly correlation2011–212021–2615 years
QQQ–TLT−0.23+0.54+0.11
QQQ–IEF−0.22+0.49+0.13
QQQ–GLD+0.13+0.09+0.10
QQQ–DBC+0.43−0.04+0.23
TLT–GLD+0.26+0.38+0.29
GLD–DBC+0.22+0.12+0.19

Before and after the flip

PortfolioSharpe 2011–21Sharpe 2021–26Max DD 2011–21Max DD 2021–26Return 2011–21Return 2021–26
Permanent1.020.46−9.2%−19.1%7.2%7.8%
Risk Parity1.360.24−5.6%−22.8%9.9%5.9%
Equity1.350.66−17.0%−32.6%22.4%16.3%
Equity Gold1.270.86−11.0%−27.0%17.0%17.7%
QQQ Top 101.350.73−21.3%−40.8%25.3%20.7%

Where the risk comes from

Each holding’s share of portfolio variance (weight × covariance with the portfolio, from monthly returns) next to its share of capital. A negative share means the holding offset risk. Equity is all QQQ, and the top 10 changes every year, so neither is shown.

Permanent

HoldingCapitalRisk, 15YRisk, 3Y
QQQ25.0%36.9%32.9%
GLD25.0%36.5%35.9%
TLT25.0%26.4%31.1%
BIL25.0%0.2%0.1%

Risk Parity

HoldingCapitalRisk, 15YRisk, 3Y
QQQ35.0%57.8%55.6%
TLT25.0%23.1%30.1%
IEF25.0%11.2%13.3%
GLD7.5%5.8%3.2%
DBC7.5%2.1%−2.2%

Equity Gold

HoldingCapitalRisk, 15YRisk, 3Y
QQQ75.0%88.9%91.9%
GLD25.0%11.1%8.1%

The building blocks, 15 years

Each ETF on its own, total return. Sharpe against T-bills. Recovered is the first month-end back at the peak before the maximum drawdown.

ETFExposureReturnVolSharpeMax DDPeak → troughRecovered
QQQNasdaq-10020.3%17.5%1.07−32.6%Dec 2021 → Dec 2022Dec 2023
GLDGold6.0%16.3%0.35−41.0%Sep 2012 → Dec 2015Jul 2020
TLT20+ yr Treasuries−0.1%13.1%−0.06−47.6%Jul 2020 → Oct 2023Not yet
IEF7–10 yr Treasuries1.1%6.2%−0.03−23.2%Jul 2020 → Oct 2023Not yet
DBCBroad commodities2.7%16.9%0.15−61.8%Feb 2012 → Apr 2020May 2022
BIL1–3 mo T-bills (cash)1.5%0.6%—−0.4%Sep 2011 → Oct 2015Jun 2017

Sensitivity to the rebalancing rule

Fifteen-year results under other schedules; the shaded row is the quarterly rule used above. “Never” is buy and hold from September 2011, which lets QQQ drift to dominate and changes the portfolio rather than refining it.

Permanent

RuleReturnVolSharpeMax DD
Monthly7.3%7.8%0.76−19.1%
Quarterly7.4%7.8%0.77−19.1%
Annual7.6%7.8%0.80−18.6%
Never11.6%11.4%0.90−27.2%

Risk Parity

RuleReturnVolSharpeMax DD
Monthly8.3%8.7%0.80−22.9%
Quarterly8.6%8.7%0.82−22.8%
Annual8.7%8.6%0.84−22.4%
Never13.2%12.7%0.93−29.5%

Equity Gold

RuleReturnVolSharpeMax DD
Monthly17.1%14.1%1.09−26.9%
Quarterly17.2%14.1%1.10−27.0%
Annual17.3%14.1%1.11−26.8%
Never18.4%16.5%1.03−31.5%

Since the last month-end

Total return for Oct 1–16, 2026, which the windows above leave out, and the drawdown from each portfolio’s highest month-end value including this move. For the QQQ Top 10 this assumes the September 30, 2026 list matches QQQ’s top 10 on October 16.

Permanent−0.1%From peak −3.4%
Risk Parity+0.2%From peak −2.3%
Equity−4.0%From peak −4.0%
Equity Gold−3.0%From peak −4.2%
QQQ Top 10−3.6%From peak −5.0%

QQQ −4.0%GLD +0.3%TLT +3.3%IEF +1.6%DBC +4.4%BIL +0.2%

What this is anchored on

ItemDataSource
ETF pricesMonth-end dividend-adjusted closes, Sep 2011–Sep 2026, for QQQ, GLD, TLT, IEF, DBC and BILdigrin.com; checked against PortfoliosLab (2021 onward) and officialdata.org (every QQQ month)
QQQ’s holdingsEvery position and its market value at each September 30, 2011–2025Invesco QQQ Trust annual reports, SEC Form N-30B-2 (e.g. 2011, 2025)
Stock pricesMonth-end dividend- and split-adjusted closes for the 22 stocks that made the listdigrin.com; checked against chartrow.com since November 2016
Today’s holdingsQQQ’s holdings on October 16, 2026stockanalysis.com
Daily drawdownsQQQ’s drawdowns on daily closesPortfoliosLab
October 2026Month-to-date returns through October 16PortfoliosLab (ETFs); digrin.com and chartrow.com (stocks)

Method

The four fixed portfolios start at their target weights at the end of September 2011, drift with returns, and are reset to target at every quarter-end close. Cash is BIL, a 1–3 month Treasury bill ETF, net of its fee; the same series is the risk-free rate, so the cash sleeve adds return but no excess return. Every window starts on a September quarter-end, so each trailing window is the same as a fresh start on that date. There are no trading costs or taxes.

The QQQ Top 10 takes QQQ’s holdings at each September 30 from its annual report: the 10 largest positions by market value, weighted by that value scaled to 100%. Because these are QQQ’s own weights, the Nasdaq-100’s caps on its largest names carry through. Between lists the weights drift with prices, as in a cap-weighted index. Stock prices match the value per share in QQQ’s reports at every holding checkpoint. Kraft Foods, in the September 2012 list, spun off Kraft Foods Group in October 2012; the price source treats the spun-off shares as reinvested in Mondelez, which differs only slightly from holding both.

Statistics. Annualized return is geometric. Volatility is the sample standard deviation of monthly returns, annualized by √12. Sharpe is the mean monthly return over T-bills divided by its standard deviation, annualized by √12, with the standard error from Lo (2002). Sortino divides by the downside deviation of excess returns; Calmar is annualized return over the size of the maximum drawdown. Drawdowns use month-end values. Risk contributions use the sample covariance of monthly returns at target weights.

Checks. An independent re-implementation from the raw inputs reproduced the headline figures, and the stock returns agree with a second source to within a few hundredths of a percentage point a month on average. Nothing here is investment advice.