Faber 10-Month Timing — Pine Script v5 strategy

This script applies Mebane Faber's 2007 timing rule to a monthly chart. It holds a long position while the month closes above its 10-month simple moving average, and steps out when the month closes below it.

Full code (free to copy)

//@version=5
strategy("Faber 10-Month Timing", overlay=true, pyramiding=0, initial_capital=10000, default_qty_type=strategy.percent_of_equity, default_qty_value=100, commission_type=strategy.commission.percent, commission_value=0.1, process_orders_on_close=true, margin_long=0)
smaVal = ta.sma(close, 10)
if (close > smaVal)
    strategy.entry("Long", strategy.long)
if (close < smaVal)
    strategy.close("Long")

Results from July 2007, measured 31 October 2026

Lab 0.1.0-253 judge, frozen rule with no tuning, 10 bps per fill. Rule chosen and frozen on 12 October 2026. The drawdown is descriptive: the judge does not count it in its verdict. Data: Robert J. Shiller; Kenneth R. French Data Library; World Bank Commodity Price Data (CC BY 4.0). Research on past data, not advice.

* Biased rows, corrected on 31 October 2026: the S&P 500 (Shiller) and gold (World Bank) series are monthly averages, so filling at that average happens partly before the signal is known. Gold recalculated with the fill one month later gives +322.5% against about +550%, with a 30.0% drawdown. That is a cautious case: the true figure lies between the two. The S&P 500 row was not recalculated and carries the same bias. Read the US and ex-US rows first: they use month-end returns. The code also ran on TradingView on 31 October 2026, with no compile error, on monthly bars. SP:SPX from 1871: +447,809% including the open position, against +155,216% for buy and hold, with a 37.8% drawdown and 103 closed trades. TVC:GOLD from 1832: +9,790% against +18,656%, with a 43.0% drawdown and 77 closed trades. Do not read these as the same test: most of that history predates the 2007 publication, SPX excludes dividends, which penalises buy and hold, and the October 2026 bar was not closed. The judged table above is our measure.

Read the full test of nine old rules →

How the script works

The header sets //@version=5 and calls strategy() with 100% of equity per entry, 0.1% commission per fill and process_orders_on_close=true, so orders fill at the close of the bar that gives the signal. smaVal is ta.sma(close, 10), a fixed length with no setting to tune.

When the close is above smaVal, strategy.entry("Long", strategy.long) opens the long position, or does nothing if it is already open. When the close is below smaVal, strategy.close("Long") closes it. The script never goes short and holds no position between an exit and the next entry.

The rule counts bars, not calendar months. It only means 10 months on a monthly chart: on a daily chart, the same code uses a 10-day average, which is a different strategy.

Parameters in the code

Known limits of this rule

• In our test it did not beat buy and hold on any of the four markets, and the judge's verdict is inconclusive on all four.
• A monthly rule makes few trades: 10 to 19 over almost 20 years. That is too few to prove an edge.
• The rule only looks at the month's close, so a fall during the month is not cut until the month ends, and a whipsaw costs two fills.
• The script earns nothing while out of the market, and it includes no tax, spread or financing.
• An index chart on TradingView may exclude dividends, so its result will differ from the total-return series in our table.

Before you trust it: test it on data it never saw

A backtest in TradingView's Strategy Tester shows one past path with the settings you picked. Strategy Arena Lab (free, Windows and Linux) replays this Pine on your PC, compares its trades one by one with TradingView's for supported Pine, counts the variants you try, then judges the rule on a period kept sealed during the search: holds, mirage or inconclusive.

Get the free Lab →

Other free Pine Script v5 strategies

Educational code on past data. No investment advice; no result is promised.