Hey guys, This weekend, I tested an interesting idea from "Short-Term Trading Strategies That Work" by Larry Connors. The original idea doesn't work anymore. However, with some adaptations, I made it functional. The strategy uses the 2-period RSI, what Larry Connors called "the Holy Grail of indicators." Here's how it works: Equity and drawdown curves for the elaborate strategy (N = 2) At the opening of every trading session: We will split our capital into N slots and buy stocks whose 2-day RSI from the previous day closed below 5; If there are more than N stocks in the universe with the entry signal triggered, we will sort them by market cap and prioritize the smaller ones; We will hold N positions maximum at any given moment; When the stock closes above yesterday's high, we will exit on the next open. To ensure we only trade liquid stocks: We will focus on large/mega caps, skipping small caps; We will restrict ourselves to only trade stocks that have been traded in all sessions over the past 3 months from the day in question; We will only trade the stock if the allocated capital for the trade does not exceed 5% of the stock's median ADV of the past 3 months from the day in question. If we are holding a particular stock and it closes below its 200-day simple moving average, we sell at the next open. Summary of the backtest statistics Summary of the backtest trades I statistically verified the edge and its size, how it varies with market cap (which is why we sort by market cap, ascending in choosing the opportunities to buy), and tested several different setups. The full write-up is here. I'd love to hear what you guys think. Cheers
The strategy presented & demonstrated in the book trades only 1 instrument. In my tests, thinks kind of worked, but most of the results came between 1999-2010. I changed the strategy by trading several instruments in parallel instead of only one. This led time exposure from 10-12% to 90-98%, which significantly improved the returns.
It's all about the drawdown not the profits. It would be hard to stay with that system cause of the drawdown percentage and the length of time in the drawdowns. Also on the matter of drawdowns they happen very quickly and are very slow to recover. Just one other observation if I were a client, I could have been using the over 50% of my money that was tied up in those drawdowns for the past 1.3 years to be doing something else that is making money. It's like I loaned you over 50k for 486 days without any interest. This would not be my HG. But I appreciate the work you put in on it, you're thinking and exploring.
I have the usual complaints for one back testing is worthless and useless, two a 35% maximal drawdown on $100,000 account is a lot of drawdown and a maximal drawdown period of 486 days is near eternity. Where is the profit loss per trade or should I go with 30,963?
%% Sounds like a real fair way to test it,1999-2010. Also his max drawdown on SPY is 56.8%, good numbers then; that also sounds like buy + hold =56.8%. Fair enough. Most likely better if some one used something like his 200dma sell, more or less; a 50 week moving average may have helpd a bit, 2022 bear ,I checked. NOT that 2022 bear was as deep as 2008, it was not Thanks.
Agreed. Although I should add I find his (her?) posts are pretty good by ET standards since he make specific claims that can be tested. I'll go and backtest all of his strats in a year or something with new data as it should be pretty fun to see forward performance.. have a hunch it will "go like usual" with hand made price based strategies but I would like to be surprised.
RSI(2) < 5 means just buying short-term dips, no? And you believe in a version of the world where the names that recover best are the ones that have been sold the most. Which is fine. I'd like to see the performance of this strategy when spread against either simple buy-and-hold SP500 or some kind of RSI(2)>95 selection of names.
How bout adding volume to the study, if it closes high on weak volume, then exit on next close above high, otherwise hold position.
Yep, I read that Book too. I assume nost here have messed around with different RSI time periods. Sorta works using different time periods, but the RSI goes to shit when the Funnymentals change unforunately - overbought/oversold will mean nothing eg this Years SPX Bull run, the RSI stayed overbought since December.