Probably two moving averages slightly spaced apart should suffice, so on cross it determines rising or falling of the primary MA. A measure of acceleration/deacceleration would be the distance apart.
Ok well I just backtested EUR/USD to see what would happen if I bought every time the momentum of 50MA turned positive, and sold every time it went negative. I was encouraged to do this because I saw it would have gotten me a nice gain since from 5/28 - 9/22. But when I went back 9 months, I would have lost money at an annual rate of -334%/year. Believe it or not, I've been down this road before... Of mathematically analyzing past TA indicators... And have never found anything to give me a statistical edge.
I changed it so that I only buy-to-open when 50ma>200ma, and only sell-to-open when vice versa... Now I only lost money at -135%/year.
Here's a better idea for you..... Never mind your suggestion above, but try this one for going long as an example. When closing price is greater than 50SMA, AND bar length greater than 3%, AND close greater than open, BUY. Now I haven't tested it, but I bet ya get better results. A small tweak can make a big difference to results.
Funny anecdote: The trade I mentioned above that had a historical loss rate of -333%... I just closed it manually at a gain of $765.40 (on $10,000 margin). That's great money for a few hours. Although I take zero pride in that because I know it was just luck.
3% bar length.... If yesterday's close was $10, then 3% is 30c. Bar length top to bottom = 30c. Sell signal, mmmmm, ok, plucking something out of air for you, sell when price is greater than 3% and close lower than open. Never mind the MA on sell signal.
Problem using 3% on EUR/USD because that would require a daily move of about $0.035. That never happens. I want to stay with Fx pairs because I think they are most likely to follow TA patterns compared to other instruments. EUR/USD has a low bid/ask spread so it's a good choice for me.