Just to start a new thread away from GDT about how to improve the BB signals. I had noticed that some closed with open red candles (advance on the day) even though they closed under the lower BB. I wanted to see if those trades improved the odds. I found 25 trades from the NASDAQ 100 in the last 100 days. Two are still going (ATYT and BMET). Of the 23, 19 were winners (82%), 4 were losers. Average loss was 3.33% (there was one big 8%) and average win was 1.78%.
An account starting with $10 and paying $10 a trade and continuing to use only the $10K and not increasing trading size had $11894 after the 25 trades. I tried having only a subset of the trades and here are the results
only over 200 sma $10316
Exclude signals on fri $11746
only over 50 sma $10116
Less than average vol $10034
Stochastic < 30 10708
Stochastic > 30 11320
Stochastic RSI = 0 11136
So taking every trade looks the best for this subset of BB signals,
h

Howard,
I'm not sure if my thinking is logical or not but it seemed to me that we're going to get the occasional large loss no matter what filter we use so I thought that what ever filter or combination of filters that give the highest percentage of winning trades would be the way to go. I don't mind passing on trades and waiting for the odds to be a little more in my favor. You have some interesting ideas that wouldn't have occured to me. Like excluding Fridays and Stochastics. I thought I was satisfied with the 200 DMA and volume filters, but now you've got me thinking again. I've got more testing to do. Cool!

Jim

P.S. Did you mention something about 20 DMA in another post?

Yes, it was that the 20d sma today is higher than it was

Jim,
I was thinking about what you wrote. What I want is a high expectancy and that may not be the highest percentage of winning trades. If for instance a filter gave us only 60% winners but the average win was 3% and the average loss was .4%, that would be a very winning system (and if you find such a filter I'll be happy to hear about it). That's why when I model, don't just rely on percentage winners and average gain; I just plug the trades into a spreadsheet that simulates how each filter would trade the system and show the amount that say $10000 would become. Different systems will also change based on amount traded, such as the bb. If you only used $1000 or $5000 per trade you'd have a lower expectancy than if you used $10000.
just random thoughts. It was a good bb day today.
h

Howard,
If you want a high expectancy high profit system you might want to take a look at the new idea post. I ran a test on the system today taking long and short trades for the Nasdaq 100. I'm getting about 29 trades, 25 profitable, 4 losses, 86% profitable trades, average winner 8.3%, average loser 6.7%. I tossed out one very large winner that would have skewed the results higher by another percentage point. My definition of an up trend or down trend may be different than somebody elses though. The downside for the system is larger drawdowns and a longer time frame so money is tied up for a considerably longer time than the B/B system. I'm just estimating but I would say that the average holding time for a trade is somewhere around 2-3 weeks.

Jim

I'm getting an average holding period of about 28 working days per trade.

Gary