something interesting I found in manual backtesting is that buying breakouts of the high or low for the first X days in a calendar month has about a 65 percent win rate if held to close of the month.
One might be able to layer intraday acd with this 65 percent bias and only take A trades in the direction of the daily bias to enhance returns....im sure many have already looked at this but figured I would throw it out there.
One might be able to layer intraday acd with this 65 percent bias and only take A trades in the direction of the daily bias to enhance returns....im sure many have already looked at this but figured I would throw it out there.
(btw I saw his 20001 post in some thread a while ago). I also coded reversal indicators from Fisher's book and found that they very often signal reversal but needs some other confirmation (have this signal on FXI now (china) and knowing that it is a big bubble and fundamentals are not that good over there and almost 50% of that etf is chinese financials it could be a good fundamental confirmation so maybe its a good moment to try timing this bubble). I use automated NLs which are nice but not that terribly effective alone. For levels, I like to have some "story" or theory and then, I might fade weekly or monthly if it fits my theory, but if A later confirms then my idea is probably wrong. If A is confirmed I usually wait for pullback to that A level but that sometimes misses good trades. Also I have coded my application which shows prices in columns like Livermore described in his book. It is useful to quickly see where the price is - bullish trend, bearish trend, natural reaction, natural rally etc. That helps when watching a lot of products, but I will probably need to scale down on that - too much is not good either as you loose focus... My master plan is machine learning trading so currently working on that