Hi, I'm doing some work on patten recognition and have been cross referencing those algorithms against previously published candlestick patterns as a kind of blind test. One of the books I've been using for this is "Japanese Candlestick Charting techniques" by Steve Nison. In particular I've been look at the tasuki Upside Gap pattern shown on p.130 against a weekly chart of PL from 27/8/1990. And this is what brings me here. I've compared that chart with my pinnacle data and there is no tasuki pattern. Indeed, the entire price chart looks completely different. Can anyone help me out here? Is this book inaccurate? Is the chart shown in the book PL or not? if not, what is it? Otherwise, on a more general note, if anyone can direct me to good historical examples (not definitions) of candlestick patterns against which I can validate my code I'll greatly appreciate it. Thx D
Probably it was due to data adjustments that the patten was lost. Try undjusted data. The patten recognition guru in the world is Michael Harris. In one of his recent blog posts he claim that data adjustments distort patterns and cause chaos in TA and backtesting: http://www.priceactionlab.com/Blog/...ting-part-ii-more-on-close-vs-adjusted-close/ I think so too. I suggest you look for the actual front contract for that time period. Then you will know whether the pattern in the Nison book really formed.
=============== Well yes, borrow or buy William O Neil books; all of them. Even though you will have to use barcharts instead of candles, but that will work mostly....................................... Wisdom is profitable to direct; I never have really been able to use exactly his ''cup with handle'' pattern , even though its his most famous pattern That one aslo makes no difference candle or barcharts