I bumped into a book recently called "Short term trading strategies that work" by Larry Connors and Cesar Alvarez.
It's very short so I finished it in two evenings. Quite interestingly most strategies described in the book are of swing trading pattern: the main thesis is to buy a temporary pullback in an upward market.
It has some really interesting results, such as a super simple "double 7 on S&P" strategy -- buy $SPX when at 7 day low, sell when at 7 day high -- practically captured all of the index's gain over the backtesting period, while only in the market 30% of the time.
Since it's also a quite old book -- the backtesting period was 1995 - 2007, I decided to test on 2017-2026 to see if the same strategies still work today.
Well spoiler alert, they don't :) I didn't find any results that beat SPX index.
I guess I shouldn't really be surprised, any published winning strategies would have been scooped up in this crazy algo trading age. I did learn something valuable though:
* The long bull market we have been in for past decade means it's *really* difficult to beat index if you are sitting in cash most of the time.
* It also means some pullback scenarios happen less frequently which further contributed to the underperformance: most strategies had less time in market in the past decade compared to the book result
* The book has one strategy using VIX: buy SPX when there is a VIX spike, sell when it calms down. This is *very risky* in a major downturn event where elevated VIX levels persists for a long time -- in my backtest, the COVID sell off practically wiped out 30% of portfolio in one trade.
* The proliferation of algo trading erased the easy inefficiencies in the market: for example, the book showed that fund managers often buy stocks at month end as new pension/salary contribution comes in, which could result in a small edge. Well that edge no longer exists.