Trading edges are being arbitraged out of the market at the speed, well not of light, but of blogging and academic paper publishing. This necessitates identification of new, idiosyncratic and secret edges, which is a dreadful task for most of those who used in the past strategies from the public domain.
Strategies found in the public domain might have worked for some time but as their trades get crowded, it is natural that the market will not be able to provide profits to all those who believe they can copy a strategy from a book, magazine or blog article, or from an academic paper and profit.
Intraday and short-term strategies have seen the biggest and immediate adverse impact and deterioration. Next come trend-following and tactical asset allocation.
“But the professor said trend-following has worked in the last 100 × N years.” (where N is an integer as large as the academic competition for tenure can make it.)
Of course it worked in simulation, but it may not have worked had participants actually used it back then when there were no phones, cars, or even electricity.
The above answer reveals “hidden” assumptions of backtesting that some academics either do not understand or if they do, they try to obscure them.
The assumption [of backtesting] is that history repeats itself and the reaction of the participants to the same information flow is the same. The empirical shortcoming is that the trading system is not a market participant during back testing
The above quote is from my book Profitability and Systematic Trading, Chapter 6, Get your free copy here under free books.
Let us take a look at an example of a “edge” that has been having problems lately: the famous RSI2 mean-reversion long-only strategy in SPY. Note that all backtests include $0.01 per share commission and equity is fully invested.
Below is the equity curve performance from SPY inception to end of 2017 for the parameters noted on the chart.
The strategy outperforms buy and hold on a risk-adjusted basis. MAR for the strategy is 0.32 versus 0.18 for buy and hold.
Below is the monthly returns table for the above strategy:
The strategy managed a positive return in 2008 despite the bear market. Since 2014 returns are low but still positive. But this is what happened after 2017:
In 2018 the strategy generated its largest ever annual loss equal to -8.5%. Year-to-date, the strategy has not generated a single trade although SPY is up more than 8%. Here is a chart of SPY and the indicator that shows this lack of trades this year:
Someone (usually a naive trading strategy developer) could claim that a different set of parameters could have worked better in 2018. This is true: entry below RSI2 = 10 and exit above RSI2 = 90 would have produced +9% return in 2018. Why is this thinking problematic?
Below is the backtest since SPY inception to end of 2017 with the new parameters that offer positive performance in 2018:
It may be seen that these parameters would not have been selected since MAR for strategy is nearly equal to buy and hold MAR.
The conclusion is that you cannot know in advance which parameter set to use. In fact, these strategies lack intelligence in adapting to new market conditions. Yet, many based a business on this and related strategies and were not quick to terminated them but remained until the end in hope that equity will reverse to upside but losses accumulated instead.
We have shown just one example of an edge that appears to have vanished. We have written in past about others, for example the IBS. Net loss for 2018 was -12%, higher than at the time the article was written.
What to do
Public edges will diminish. Thousands of bloggers with no skin-in-the-game and finance professors in search of a few tenure credits will make sure of that. The question is what are the choices left. Advanced machine learning with features that have economic value is one of them. Using alternative data is another but signal to noise ratio is very low. Going back to discretionary trading is another choice for the few that have lived through at least one bear market and know how it feels.
Those who are insisting in using pedestrian strategies are probably doomed.
If you found this article interesting, I invite you follow this blog via any of these methods below:
If you have any questions or comments, happy to connect on Twitter: @mikeharrisNY