RE #6 You specifically mentioned large cap though. I agree that higher volatility is "better" only because that's my personal preference. Small/micro cap stocks might get excluded in my backtesting because I exclude when bid-ask spread is too high. Otherwise my systems don't discriminate based on market cap, whether high or low. RE #4 As mentioned above in this thread, ETF's are a different animal, so it's usual to have an "only equities" and "only ETF's" stragegies. However I wouldn't recommend ignoring commodities completely. Your system might not trade them, but they do affect the market. Metrics I don't keep track of, other than did the system perform as well in real time walking forward as predicted in backtesting. If not, back to the drawing board!
Keep in mind there is at least one known regime shift, I forget the year, but within the last 20 years for sure and possibly within the last 10. So testing for 20 years may be too long. (Depending on the strategy. It is possible to have a strategy is regime independent and in that case 20 years would be fine.)
I have been ruling out the Chinese small caps recently. Too many frauds lately. I'm not trading ETFs or large caps because the opportunities are better in small cap stocks.
I can see your reasoning with Chinese small caps. Did you consider excluding a subset of ETFs, such as commodity related ones (which have performed poorly in my backtests). According to Larry Connors' research, ETFs have higher win rates than stocks. If you are craving more volatility, there are the 3x ETFs, although there is not as much historical data for them to test with. How about excluding any low volatility stocks as opposed to all large caps?
I shouldn't have said large caps. I exclude based on low volatility, and it just happens that a lot of large caps fall into that set.
I'm a fan of Larry Connors research, but he makes too much of the win rate. His Daily Battle Plan wins something like over 80% of trades but after costs his clients are still likely to lose money: http://www.cxoadvisory.com/individual-gurus/review-of-larry-connors-daily-battle-plan/ I've done some testing of ETF's and it is worthwhile to look at commodities funds differently, but they are still opportunities ... I just like the opportunities in stocks better.
Hi Mark, I perform nightly scans to generate a "watch list" for the next trading day in the rtm automated system I run. I exclude any stocks with margin restrictions, so I can maximize my buying power. This generally means no OTC/Pinksheet stocks and no 3x ETFs. I am with you on no Chinese pump and dumps. Most of the time, ETFs are not volatile enough to generate signals for me anyway or if they do, then % gains are generally not great. I just take the risk on news related events adversely affecting me. I figure I will be stuck with a "statistical outlier" every few months, so I only take 100-300 share positions. I figure news related event have just as much chance of helping me, so I view these outliers as just cancelling each other out within my overall profit distribution. Recently had VSEA jump 50% for me on a buyout. Essentially, if a stock is not volatile enough or trades with too little volume, it will exclude itself from generating a signal within my system.
Why no ADRs? Are we jus talking FX move issues [hedgeable]. Also, why is a wide spread an issue .. if you're receiving it
Rol, how many max concurrent positions per portfolio do you take to diversify away news related risk? Do you try to diversify individual accounts that you trade in, so if one of your accounts is relatively small compared to others, how many positions would you divide it into? Does everyone else ignore news? How about trading around earnings, guidance revisions, secondary offerings, shorting announced mergers, shorting companies reported to have hired someone to find a buyer for self, etc?