I was really interested when i read about this. http://groups.google.com/group/misc...ad/31ea4ac4aea16e99/?#doc_a2ee6622d3d84c57... hopefully that'll be an active link.... at any rate. hmmm i really don't know how to begin this. i'll post some annotated charts later in the day to get some feedback on if i'm doing it just right. i'm currently going through the 2nd level, and drilling that into my brain.
Will be interested to see what the backtesters have to say. Should be fairly easy to code the parameters. Seems like it would work well with large swings but not too well in a low volatility environment. This method was posted in 2000 when things were a bit different.
ok lets try this out, i went over 9/23 (friday). this is what i came up with. note: i'm leaving starting at 10:00am and ending at 3:30pm est my testing showed significant improvement doing this. comments regarding anything posted are welcome. did i do anything 'wrong' with this chart?
side note: from what i've looked at so far with other day's charts....specifically strong trends, the 2nd level (as i'm doing it, which may be misunderstood on my part) seems to do really lousy reason being that the rising trendline coming off the sloped bars gets you out and has you reverse (causing small profit) then since you are countertrend now in a strong trend you get whipped around for another nasty loss before setting back up with the trend again.... this seems to happend a LOT on strong trends.
easy, i'm backtesting it currently, i'm 1 month in the backtest (well a bit over that, 29 days in), it's up around 4k (1 ct, mini dow, 10am-3:30pm starting from 12/18/02 doing the 2nd level...as grob states, levels 3 and 4 are better.
For Dan, I had suggested to him in a PM in addition to the post that the restrictions be put on risk. There are 2 contexts that have to be considered also, 1 being that the post with regards to DJ futures and the other being the formerly mentioned condition (ie. RISK). Since RISK and PACE are inversely proportional, the recommendation was to ZERO out the risk by executing in the most favorable pace conditions (ie. FOMC meetings). Reason being is that the tape is just running fluidly (ie. lack of stalls, hitches, dips) and super high octane jet fuel (ie. SUSTAINED EXTRAORDINARY volume). As a result, the bars just tape along. The problems most backtesters have is the silver bullet grail stuff. 1 GOLDEN RULE that is a sure winning BET... It is a limiting mentality since the fact remains that the market has a range of varying dynamics and characterstics, the two commonly phrased ones being whipsaw and trends. Thus, the backtesting ordeal winds up being applying the same analysis across the entire spectrum of pace considerations. How can this work??? We already know the answer, it is a refinement task that is iterative... Iterate and refine to slower and slower paces... Regards, MAK! FWIW... A few notes because I always make an effort to contribute in a productive manner, you may note that stretching the vertical axis and horizontal axis also affects results due to the usual equation for drawing a line... Nonetheless, keep up the great work...
"risk and pace are inversely related" where pace is high, risk is low? "iterate and refine to slower and slower paces" by this do you mean, trade only in say the 9:30-noon...then 1:30 to close periods? (i notice that volume and price movement drops off noticeably during the noon to 1:30 period). by iterate, do you mean the iterations of the strategy? slope pairs, 30 min bar in between, and retrace value? i'm having to get a handle on some of these words and meanings. pace especially.
Could you post the backtesting algorithm used for your tests? Feel free to PM if you do not want to post it publicly. - Spydertrader
sure (if i understand what you mean). 1. B/O of prior bar hi/low (take trade in that direction) 2. draw line from prev. consecutive bar lows (if long) or bar highs (if short)....projecting into current bar formation. 3. hold trade until/unless the current bar breaks the TL....when it does, reverse position (if no sloping TL can be drawn, use horizontal TL as reverse point...horizontal TL is hi/low of prev. bar) that is the 2nd level of refinement as described by grob (as i understand it). the 3rd and 4th levels i'm a bit clueless on.
Hi Mak, Here is how it worked over the last two years of fed meetings starting only at 2:15 pm EST. 8/9/05 -137.50 6/30/05 137.50 5/3/05 -962.50 3/22/05 837.50 2/2/05 -125.00 12/14/04 -200 11/10/04 -162.50 9/21/04 250 8/10/04 -37.50 6/30/04 275.00 5/4/04 -50 3/16/04 -537.50 1/28/04 825.00 12/9/03 50 10/28/03 137.50 9/16/03 -12.50 8/12/03 -137.50 6/25/03 637.50 5/6/03 -437.50 3/18/03 75 1/29/03 -462.50 The net BEFORE commission is 232.50 and after commission is a loss. This doesnt seem the stuff of $75,000 per year per contract does it ?