Twenty S&P 500 model portfolios: five factor strategies —multi-factor, momentum, value, quality and growth— in four sizes, each with its full backtest since 1998. It opens on the 20-stock multi-factor portfolio, the one that combines all four factors. The holdings are today's; the track record is measured with a monthly rebalance and without commissions or spread, so trading these lists every day is not the strategy that was tested.
Strategy
Holdings
S&P 500 Value Index (Top 10) Updated: Oct 23, 2026
Institutional replica of the S&P 500 Enhanced Value Index (SP500EVP). Each company's Value Score combines three fundamental cheapness ratios: Earnings Yield (E/P), Book Yield (B/P) and Sales Yield (S/P). Value Score = mean(Z_EY, Z_BY, Z_SY). The Top 10 are bought EQUAL-WEIGHTED (10.00% each), which is exactly the portfolio the walk-forward backtest above this table measures.
How the stocks are picked: model rules
Universe: the ~500 companies of the S&P 500.
Earnings Yield (EY) = trailing-twelve-month EPS / price. How much profit the company earns per dollar of price. Higher EY = cheaper relative to earnings.
Book Yield (BY) = 1 / Price-to-Book. The discount to the book value of equity. Higher BY = cheaper relative to tangible assets.
Sales Yield (SY) = 1 / Price-to-Sales. The discount to revenue. Higher SY = cheaper relative to its ability to generate sales.
Each metric is normalized to a Z-Score (μ=0, σ=1) and winsorized to [−3, +3]. At least 2 of the 3 metrics must be available.
Value Score = arithmetic mean of the available Z-Scores: (Z_EY + Z_BY + Z_SY) / n.
Weighting: EQUAL-WEIGHT. Each of the 10 positions weighs 10.00% and the weights add up to exactly 100.00%.
Why equal-weight and not market cap: the curve and risk metrics above come from an equal-weight walk-forward backtest, so the weights in this table must be the same or the chart would describe a different portfolio from the one you are being shown.
Selection: Top 10 by Value Z-Score.
Total Return:
Recent days today's holdings, through the close of Oct 23
Walk-forward backtested rule
1 Day:-1.47%
1 Mo.:6.20%
2 Days:2.33%
3 Mo.:6.48%
3 Days:2.11%
6 Mo.:11.30%
4 Days:3.30%
YTD:27.85%
5 Days:2.70%
1 Year:26.88%
Two different things, so they are in two columns. The left one is the portfolio the model picks today, priced over the last few sessions. The right one is the rule run forward through history, rebalancing monthly. Over one to five days a 10-stock portfolio moves mostly on noise; the annualized figures are in the box beside this one.
Risk of this model portfolio (simulated) · full history:
CAGR (Annual):13.92%
Max Drawdown:-82.20%
From 1999-02-02 to 2026-10-21 · 27.72 years
More risk ratios: Sharpe, Sortino, Calmar, volatility…
Sharpe Ratio:0.50, weak
Sortino Ratio:0.73, weak
Calmar Ratio:0.17, weak
Volatility:32.11%
Beta (vs equal-weight, monthly):1.61, middling
Alpha (Jensen):-1.30%, weak
Treynor Ratio:0.07, weak
Monthly Win Rate:60.06%, middling
How these figures are calculated
Walk-forward with point-in-time data: at each date the ranking uses only prices up to that day and fundamentals already filed with the SEC, the top 10 are bought equally weighted, and the portfolio is rebalanced monthly. 334 rebalances between 1998-12-31 and 2026-10-21. The universe is point-in-time: at each date only companies that were in the S&P 500 that day can be selected.
Limitations you should know about
Total return: dividends are reinvested.
Commissions and bid-ask spread are not included.
The universe is point-in-time: at each date only companies that were in the S&P 500 that day are eligible. Up to 2026-08-21 the history comes from Sharadar, with the prices of those that later left the index (about 0.1% of member-slots are missing); since then, each day's data.
Quarterly (TTM) fundamentals filed with the SEC, used only from the date they were actually published.
Download this curve as CSV — every rebalance, the portfolio and the equal-weight S&P 500, from the same file this page reads.
Ticker
Sector
Last Price
Today live
1 Wk Ret.
1 Mo Ret.
3 Mo Ret.
Z-Score
Algorithm Weight
The 10 stocks S&P 500 Value Index (Top 10) holds
bought on October 21, 2026 · 2 of 21 sessions · 19 left until the next rebalance
Cumulative total return, dividends reinvested, rebalanced monthly. The band is the distance to the index.
Difference +2,249.76 pp
-1000%0%+1000%+2000%+3000%+4000%
Fall from previous peak
-100%-50%0%Feb 99Aug 04Mar 10Sep 15Apr 21Oct 26
From February 2, 1999. Worst fall: -79.40% (peak Jun 11, 2007, 21 months to fall; trough Mar 11, 2009; back to peak Jan 7, 2011, 22 months to recover). At the end of the track record it sits 3.45% below its peak. The bottom of this chart (-79.40%) is shallower than the published max drawdown (-82.20%), which is measured on daily closes: this chart has one point per month, so anything that fell and recovered within the month does not show. It sat below its previous peak 71% of the track record, against 63% for the index measured on the same points. The longest stretch without a new peak lasted 43 months, from Jun 11, 2007 to Jan 7, 2011.
Year by year whole track record, does not follow the selector
Beats the index in 13 of 28 years. The best full year was 2009 (+104.58% against +45.91% for the index) and the worst 2008 (-54.90% against -43.61%). Years marked * are partial: the first one starts where the track record starts, and the last one is the year in progress.
All of this is measured on the monthly points of the chart; the Max Drawdown in the risk box uses daily closes and is therefore deeper — sometimes much deeper: a fall that recovers within the month leaves no trace in monthly points. The drawdown panel footer gives both figures.
How these figures are measured, exactly
The fall from peak compares each monthly point with the previous high of the series. Since a deeper low may have happened between two monthly points without showing here, this curve bottoms out higher than the Max Drawdown in the risk box, which is computed on daily closes.
Year by year is measured between December points, because monthly data has no 31st close. That is why the year in progress does not match the YTD figure to the decimal, and why the first and last years carry an asterisk: they are parts of a year, not whole years.
Entries and exits are the difference with the previous rebalance, dated September 22, 2026. They are decided by the published rule; this is not investment advice.
The twenty portfolios, one by one
Next to each one, its annual return and its worst fall, both from the full backtest and exactly as carteras.json publishes them. They do not rank the same way, and the first column should not be read alone: on the long history, the top earner is also the one that falls furthest.
Value 10-stock S&P 500 portfolio
The Value 10-stock portfolio splits its money across the 10 S&P 500 companies with the best value score (how cheap they trade against earnings, book value and sales), in equal parts: each holding weighs 10.00% and it rebalances once a month. With today's holdings, the 10 are spread across 7 sectors; the largest are Communication Services, Consumer Cyclical and Technology, with 2 each. In the simulated track record, from December 1998 to October 2026, the rule would have returned 13.92% a year with a maximum drawdown of 82.20%; the equal-weight S&P 500 computed by the engine, this site's benchmark, returned 10.22% a year and fell as far as 60.41% over the same window: 3.70 points a year more, with a drawdown 21.79 points deeper. These are simulated results, before commissions and spreads, and not investment advice.
Twenty portfolios built with the same rules as the ranking - five strategies by four sizes - holdings in the open. Here, the last twelve months; the full track record, risk metrics and methodology for each are on its card, below. Past performance does not guarantee future results.
We save the portfolios you design, your own portfolio and your Academy progress, synced across devices.
Free, no card. Two automatic emails and no more: the one that confirms your address and the welcome one. After that, nothing you do not switch on yourself in your profile — such as the alert for insider buys, the same day they are reported to the SEC.