Very intresting predictor you have in I1. On your concepts page this predictor is outlined together with SP all way back to 1965. My question is what the lag is between the data making up the predictor and the I1 datapoint in the chart. Is the lag constant for allt datapoints?
There is no lag. I run the build engine once a week and it constructs data points with precision for approximately 90 days, with ever-increasing reliance on internal extrapolation for dates beyond that. So, when I build an I1 chart it is using canned data, not relying on market data.
OK, I'm trying to figure out how it works. So if you recalculate a certain datapoint in the i1, say 12 times, during a 90 day perion (once a week) with continously shorter duration from your latest datapoints. Then I wonder what the variance is in this datapoint over time? Do you have that kind of statistics?
From your answer I understand that the i1 is calculated with no lag in the concepts page, call it i1(0). It would be very intresting to compare this dataseries with i1(90), i.e. a 90 day lag just to get an idea of how stable/independent of lag i1(x) is.
By the way, your indicator looks very promising. I just don't understand why you publish it rather than just living the easy life collecting money from us other poor souls.
I can provide a year back for you to play with. The simplest way to achieve what you want is to apply various moving averages, centered and lagged. Is that correct?
That would be nice. What I want is the i1 predictions for a certain date given from say 90 days before to the same day. So, how do I contact you? I cant find any email on your site.
Pontus I'll post I1 data values from 2007 to end of November of this year. The I1 rule set has thresholds that influence the action taken, however. These rules should be accounted for.
Hi Steven,
ReplyDeleteVery intresting predictor you have in I1. On your concepts page this predictor is outlined together with SP all way back to 1965. My question is what the lag is between the data making up the predictor and the I1 datapoint in the chart. Is the lag constant for allt datapoints?
There is no lag. I run the build engine once a week and it constructs data points with precision for approximately 90 days, with ever-increasing reliance on internal extrapolation for dates beyond that. So, when I build an I1 chart it is using canned data, not relying on market data.
ReplyDeleteOK, I'm trying to figure out how it works. So if you recalculate a certain datapoint in the i1, say 12 times, during a 90 day perion (once a week) with continously shorter duration from your latest datapoints. Then I wonder what the variance is in this datapoint over time? Do you have that kind of statistics?
ReplyDeleteFrom your answer I understand that the i1 is calculated with no lag in the concepts page, call it i1(0). It would be very intresting to compare this dataseries with i1(90), i.e. a 90 day lag just to get an idea of how stable/independent of lag i1(x) is.
By the way, your indicator looks very promising. I just don't understand why you publish it rather than just living the easy life collecting money from us other poor souls.
I can provide a year back for you to play with. The simplest way to achieve what you want is to apply various moving averages, centered and lagged. Is that correct?
ReplyDeleteThat would be nice. What I want is the i1 predictions for a certain date given from say 90 days before to the same day. So, how do I contact you? I cant find any email on your site.
ReplyDeletePontus
ReplyDeleteI'll post I1 data values from 2007 to end of November of this year. The I1 rule set has thresholds that influence the action taken, however. These rules should be accounted for.