comments 4/21/06 (Friday)

D bracket just closed: Today's D bracket just dips into the VA. On my charts it shows by 1 tick on The Ranking Page by 2 ticks. 3 of the top 5 ranked profiles (1, 2, 5) show this as well. All 3 show that the high of the D bracket is also the HOD.

All of the ranked profiles show a narrower range than usual. Remember that average range is around 11 points and currently the max range in the ranked profiles is 43 ticks which is the average range.

(As I type this the E bracket makes a new HOD but that information was not available at the end of the D bracket so I will leave the above comments as they are.)

Interestingly all the ranked profiles show the same pattern/prediction for the E bracket: The E bracket overlaps the D bracket to the downside and trades into the VA - prices attracted to the VA like a moth to light.

At the close of the E bracket we have a completely new set of ranked profiles. I can tell this without looking at the dates because all of the ranked profiles have "outside" E brackets and all of the previous ranked profiles have overlapping down E brackets.

This is what we expect because the E bracket is now overlapping. Also note that this must be a fairly common type of day because the rank % is high. There is however, a large difference between #1 rank and #5 rank; 80.2% and 72.6% which is unusual. This indicates a good match at #1.

An aside: Yesterday we had low rank %'s which indicated that there were few/no days in the database that matched the developing day's profile. This could tell us 2 things: 1. This is an unusual day type that we have not (or very rarely) seen before. 2. Don't trade this day type because we are not "trained" for this day type.
The corollary is that today is a typical day so is safer to trade because it is more predictive.
Is there any use in this tool in excluding day types (profiles/structures) that we have not seen before and by doing so we exclude high risk days and improve profitability?
A way to test this would be to look back through trading records and see if the top rank% after each bracket was high or not. If your trading results showed good profits on days that had high rank% (and losses on days with low rank%) then that would imply that predictive days yielded greater results. This would be a way to filter out dangerous days.

Back to the profiles: They are "saying": Higher probability of RE to downside and upside RE risk will happen in late afternoon if we see it.

Here is the Rank% for the #1 ranked profile throughout the day at the close of each bracket. In light of what I said earlier it might prove useful to keep a track of these percentages and/or work out what the norm is and this may help the predictive value of previous days.

D 88.2
E 80.2
F 54.9
G 56.3
H 66.2
I 65.4
J 65.2
K 66.3
L 67.7
M 64.7
N 61.5
P 60.3
Q 58.9

Although I've manually entered and tracked these figures here, I will hopefully automate this so that any 1 day can be picked out and this information will be machine generated.