Forecasting with h2o5

FLIR2 contains about four relatively simple methods for forecasting future inflections (synthetic trendlines, tau/iota valuation, h2o5, and FLIR1), as well as assorted trading strategies, and methods for determining support/resistance and trends. It's 118 pages plus a large appendix, more than 25,000 simple words, 70 charts and illustrations, is fully printable, includes as appendices one simple tutorial on using spreadsheets and one on using msPaint. It also contains TOS code (thinkorswim) for implementing h2o5 and its histograph. Literally, a 10-year-old could use this stuff.  orangequant.com

.
today, i want to talk about one issue using h2o5. no, i'm not going to show how h2o5 is derived- that info is in the FLIR2 eBook. this post mainly helps those who already have the book. but it also shows others how easy it is to use h2o5 (Hudson 2nd-order Oscillator).
.
the issue i want to address is accuracy. the longer your trendlines on h2o5, the stronger the inflection might be. the shorter your trendlines on h2o5, the more accurate your forecast. this is a purely mechanical issue- not a data issue or system failure. later, i plan to blog a trig function method to reduce such error and an AUTOMATIC calculator so you won't have to know anything about trigonometry to use it. alright, here's the problem: when drawing a line from Point A to Point B, your mechanical skills are much less important to accuracy if Line AB is very short than if Line AB is very long. in other words, a couple degrees of error on a pair of short intersecting lines hardly makes any difference; but the longer the lines are, the more that error gets magnified. this can result in forecasting the wrong date or time of inflection. here's an example:
.

.

As you can see above, the slight difference in alignment of the dark blue trendline gives you an inflection that will be off by at least one period compared to the other intersect. So you could be off a day, an hour, a week, or a month, depending on your chart's timeframe.  the best remedy for this, since you know the exact values of A,B,C,D, is algebra or trigonometry.  but that's 'way too much for most people.  so, as i said, i'll soon be doing a blog on how to use an automatic calculator to find the exact intersects.  meanwhile, the best solution is consistency in your drawing habits, i.e., always draw right through ABCD or always just above.  then backtest your method--- it should show consistent accuracy, or it should show consistent delay or advance and you would factor that in.