I'll have to take a rain check gentlemen. Here in the western hemisphere (USA) it is about 3:30am and the ladies have disappeared. I am left with a very good bottle of sherry and the promise of a sunny day tomorrow. Wishing you good luck at the bar. Lefty
Before I retire for the night, all the frantic graphing made me overlook a misunderstanding between spike and Lefty. spike, Lefty's comment above was not directed at you but rather was a reply to one of my posts. It just so happened that he posted it immediately after yours.
It doesn't get better than this, right? So I'm awaiting your algorithm for exploiting such a market, ignoring slippage and commissions. -- Mr Subliminal Given that your price curves are not composed of OHLC prices but just a single value, why not use Perry Kaufman's adaptive moving average? http://www.traders.com/Documentation/FEEDbk_docs/Archive/0398/TradersTips/Tips9803.html I have implemented this in Excel to track the equity indices each day. It works quite well. Happy Easter
Very interesting link.. and great thread! Hope the moderators don't do something retarded and close this thread due to length. Actually read every post so far, and been learning a lot!
Bruce, I like your idea of using an adaptive technique. Even though it was not my original intent, I'm prepared to do the grunt work to test KAMA on my algorithm. However, before I jump in, could you please tell me exactly what it is I should do? I see references to byproducts of KAMA in your link and am not sure if they are required. If it's any help, I have Kaufman's Trading System and Methods (Third Edition) at home as well. Sooooo, please give me a list of comprehensive and idiot-proof instructions and I shall try implementing them.
Sure. Let me think this thru...the problem as I see it is I do all my work in Excel. If you are pasting the results of your algorithm into Excel for charting then the thing to do would be to send you my Excel ss with tweaks. You can then adjust the parameters to fit your output. I know this sounds like curve-fitting...it is. But I'm fairly excited about the results of the runs analysis I've performed on the equity indices. Generally, trends, either long or short average about 7 days for the Dow and the Russell. Can't remember what they are for the other two off the top of my head. What this does is eliminate one moving average from the two-moving average system acrary described as eminating from JWH. So what is left is simply crossing and re-crossing the KAMA and putting on the appropriate trend trade. Thanks for your interest, I will get something together this evening.
Bruce, I really didn't want to create any extra work for you. If all I have to do "is simply crossing and re-crossing the KAMA and putting on the appropriate trend trade", then, using the snippet of code on p. 438 of TS&M, I believe I have all that I need. So, I'll be using 2 and 30 for fastest and slowest values repectively. As for the time period for ER, Kaufman recommends 8 to 10 periods - I'll use 10. Unless you say otherwise? Clearly I have other things going on in my life (I have no life), so all this will take some time. In the meantime I appreciate your offer, just please don't spend too much time on it
Mr Sub, All that is on the other computer, it's not on right now (energy efficient) so I can't give you a setting off the top of my head...I think the original code is for 8 days but I have made some modifications to that as well as the Alpha number for the moving average. I'm thinking I made it permanently adjustable from one cell to ease the testing. It really isn't extra work when you consider what has gone before. The reason for eliminating the second average is the lag is just unacceptable when looking at the indices and their price action. I'm thinking the system was probably developed for the commodities in particular.