Are you saying that the patterns on the charts are not predictable (a complete joke) or that discretionary/intuitively trying to guess based on a pattern is a joke? To me, these two ain't quite the same. I think the problem with most chart pattern recognition is that it's too damn subjective and you can't simply sit around interpreting stuff in a live environment. This might explain why 'themickey' has success with EOD patterns. Very interesting. Thank you! Are you willing to explain more about this chart? How it's generated and what it contains? Are these intraday highs and lows? If so, which...? If not interested in sharing, that's also okay. I tend to prefer some secrecy myself on my own research...
They are just inside and outside bars, and three day rising or falling bars, common bars people look for. Just an example not something I use. Regarding breakouts, you will need to learn some programming to do what you are talking about. I do something like that in Python and then write the results to excel.
Yes. Subjectively assigning probability to things like triangles, double/triple/V bottoms, head and shoulders, euphoria/capitulation is a dead end.
This can be quantified and automated without looking at a chart. A positive expectancy would be dependent on how creative you are with the math, and what and when you trade.
Thanks. Any thoughts on the logic...? The problem for me is that a range typically ain't fixed and might have a long or short duration. Also, there might be some 'fakeouts' that doesn't constitute a real breakout, IMO. I've been meaning to learn Python, but haven't got around to it yet as my plate is full as of now. Is it a steep learning curve to be able to do useful stuff in Python? I have an engineering background, so I have some familiarity with maths and very very basic syntax in Matlab.
I can't discuss the process publicly in detail, but volatility based studies would be a good starting point.
You will have to quantify what you are looking for. You can use a highest x lookback period, or a high/low that meets a certain volatility requirement, or some combination. You can code that stuff and then test to see if certain factors lead to greater predictive value. For confirmation you could require x minutes, x number of closes, x volume, same breakout in some other correlated market, off the top of my head. Only speaking for myself, python is very hard to learn. I've never been too computer savvy and it sounds like you are better suited to programming than me. If I can figure it out I'm sure you can too. Good luck and I recommend you check out globalarbtrader's journal here. He has some code for breakouts you can look at on his blog.
Basically, I'm just interested in quantifying a trading range, then observing what happens when the market trades outside that range. It's challenging to me since a trading range ain't always 'clean' and it might last for 2 trading sessions or 10 trading sessions. It might be 'normal' volatility or contraction of volatility. So, I'm hard pressed to find a logic that can consistently describe/quantify a trading range. Testing the actual breakout shouldn't be so hard as soon as quantifying the range is done. Thanks for the recommendations and advice on Python. I won't bore everyone with my schedule, but I have a demanding full time job and find that what's left of my sparetime is not more than I need for trading and analyzing markets. If I were to learn programming efficiently, I'm not sure if I would have much time left for the markets. But it's on my to-do list for sure...For now, I've been hiring a programmer once in a while when it's stuff I can't seem to do myself