Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Strategies & Market Trends : Dunnigan's Swing Formulas -- Ignore unavailable to you. Want to Upgrade?

To: KymarFye who wrote (20)	1/27/2002 2:16:57 PM
From: KymarFye	Read Replies (1) \| Respond to of 37

also, o) C - O < MinList(C[1] - O[1], C[2] - 0[2], C[3] - 0[3], C[4] - O[4], C[5] - O[5]) p) absvalue C - C[1] < MinList(absvalue(C[1] - C[2], absvalue(C[2] - C[3]), absvalue(C[3] - C[4]), absvalue[C[5] - C[6])) I suppose both of those could be written in more flexible ways to allow testing of diverse measuring periods... Seems that when you get a narrow range doji mostly but not completely within a topping range, at the top of a range, ways to express its significant insignificance proliferate... I'm not sure how you've been accounting or choosing not to account for relationships to "major" swing highs and lows, as this bar also happened to enhance the appearance of an incipient double top on the NDX - though (straying from the main subject) I guess it didn't turn into a 2B until 1/18, and may, according to your plots (as well as relationship to historical s/r levels) already be back into narrow range whipsaw-land.

To: KymarFye who wrote (20)	1/27/2002 7:29:56 PM
From: MechanicalMethod	Read Replies (1) \| Respond to of 37

Hi Kymar, All your suggestions are valid measuring techniques that fall short as stand alone systems but are ideal measures for creating the exception rules I need to exclude weak reversal bars. Which is just what I need so I appreciate the ideas! I have a keen sense of what works and doesn't and these are valid measures. I might do a summation of a long list of measures if I can't settle on a few that work best. A summation might give the best overall weighting. I'll put them all in a matrix and go through 1 by 1 and analyze their strengths and weaknesses for every questionable reversal bar and come up with a list of the strongest, perhaps several lists and then backtest each variation. I used to have this idea for assigning a value to each condition and then running an optimization, not a parameter optimization but 1 that chooses which combination of rules gives the best result. I never worked out how to do that. Essentialy the optimization would determine which rules were strong and which were weak ... within the limited context of converting weak reversal bars back to continuation bars. If you've any thoughts on how an optimization like that might be accomplished it would be interesting. Thanks again and best regards, MM