here is a site that has a great free screener, stay away from the illiquid names for me anything trading less than 1 miilion shares a day. There is some great stuff in here from fundamental to tech analysis, plus its free. here is a link to the patterns. http://finviz.com/screener.ashx?v=210&s=ta_p_doublebottom
Over trading has sunk more ships for daytraders then anything else. Remember less is more. Fisher talks about the concept of "next". Meaning trade A does not work out, say next, and move to trade B. Don't dwell on one stock. As a daytrader, stock selection is so important. Much more important then entry's and exists and stop levels. If you are in the right stocks, it really doesn't matter where you get in and out, you're going to make money. Do not keep trading the same stock over and over again. What I like to do is use a stop that is wide enough so that if I get stopped out, it actually repudiates the original signal. Therefore leaving no possibility of taking another trade in the same direction. That makes things very easy.
Yea I use finviz and it is great...do you happen to know of any EOD scanners that scan for those toby crabel patterns? Like NR4, NR7...this seems like a good substitute for a narrow pivot scan(don't know if that even exists)
acdtrader, I calculated the stats for the ES (S&P500 future), the results of the distribution are actually not much more than you did, just for each bar during the day.The ES is open for 6h 45min (RTH), so this gives 81 bars when using 5min bars, now I calculated the probability of making the high or low of the day after 5min, after 10min, etc., after 81 bars I know with certainty the daily high and low, so the probability is 1. The two plots at the top show the same, just for high and low isolated. Bottom left shows the values for high AND low combined (different to bottom right, which is high OR low). So the chance of having made the high and low at the first bar are negligible small, even after 40 bars - which is roughly half of the day - the probability of having made both the high and low are below 30%. By implication this means that the likelihood that we are going to see another extreme is rather high at this point. The red lines for all 4 plots show how the probability looks like when using random data and you can see that the empirical values are quite different from random walk. Regarding software, for quick backtesting / prototypes I often use tradesignal, which is - similar to Multicharts - a tradestation clone. For more serious stuff we have written a full library of code for testing and trading - mostly in Matlab.
I don't know how good your programming skills are, but when looking at EOD data you maybe could even write something on your own. For daily data you could use CSI which is a cheap reliable data source (they are also the data provider for Yahoo and Google) and the advantage of such a solution is that you have full control over everything, can define your own patterns etc.