@vince.quant: How do you know a market is trending without forecasting anything? The EMA crossover holds a position from the sign of a fast exponential average minus a slow one. Unroll the two averages and that gap is a weighted sum of past returns, so the rule is a momentum measure, and Levine and Pedersen (2016) showed crossover strategies capture largely the same premium as time-series momentum. The evidence starts with Cowles (1933), whose entire abstract read "It is doubtful." Brock, Lakonishok and LeBaron (1992) then found moving-average rules predictive on the Dow from 1897 to 1986, Sullivan, Timmermann and White (1999) corrected that test for data snooping and found no edge left out of sample, and Hurst, Ooi and Pedersen (2017) took the same object to portfolio scale as time-series momentum across 67 markets: positive in every decade since 1880 and up in 8 of the 10 worst 60/40 drawdowns, in that backtest. Limitations: the single-chart edge decayed after 1986, a decay confirmed on 1987-2013 data; the 2017 figures are backtests, with per-market Sharpe ratios reported gross of costs (the paper also gives net estimates); and the strategy's stated weak spot is fast crashes like 1987, with drawdowns of its own reaching about 25% in that sample. Research tool, not investment advice. Sources: Cowles (1933), Econometrica 1(3). Brock, Lakonishok, LeBaron (1992), JoF 47(5). Sullivan, Timmermann, White (1999), JoF 54(5). Hurst, Ooi, Pedersen (2017), JPM 44(1). Levine, Pedersen (2016), FAJ 72(3). #finance #quant #trading #algotrading #stocks

vince.quant

Open In TikTok:

Region: FR

Wednesday 15 July 2026 22:57:26 GMT