Are pivots useful ? a statistical response

fortuna, good contribution. now i would be specifically interested what pivot points turned out to be the day low/high on the S&P most often. I suspect that they're simply the H and L of previous day :)))))
 
0.10% of the high/low 19% of the time?

Let's see - 0.1% sounds very tight, but that means within 3-3 1/2 points of the high/low (CAC40) only 19% of the time

And it's not really 0.1% (because accuracy shouldn't be measured based on index price but rather intraday range) - so 3-3 1/2 points would be anywhere from about 3-10% or more of the intraday range - which isn't really very accurate at all.

Being within 3-10% of the daily range from the high or low of the day using any of 5 numbers only 19% isn't especially significant (or useful).

Try picking 2 random #s each that are RND*(Previous Day's Range) above and below the Previous Day's H+L+C/3 - wouldn't be surprising if you come out pretty similarly.

If simple HLC pivot points worked consistently and reliably (i.e., gave a material edge), they wouldn't have been dumped by the majority of pros.

Good luck.
 
0.10% of the high/low 19% of the time?
More...
The first thing is the numbers were never generated for use on the indexes. That is like trying to drive a nail with a screwdriver and then saying it doesn't work very well. Well no kidding.

Even so, those numbers above are saying that almost 1 in 5 times the projected high or low and the actual high or low are within 1/10th of 1 percent!!!!! That seems pretty good to me. If I relax the requirements a bit to say.... 5 percent of the high or low, what percentage of times does that equal? How about 10%?

Again, all you have to do is look at a chart to see that the high or low from day to day is usually not within 0.1% etc.
 
Quote from inandlong:
Even so, those numbers above are saying that almost 1 in 5 times the projected high or low and the actual high or low are within 1/10th of 1 percent!!!!! That seems pretty good to me.
More...
The point is that it is NOT very good. .1% on that index is 3-4 points. That's 3-10% of the trading range for any given day but even then only 1/5th of the time. That's insignificant. You might as well generate random #s above and below the average - you're just as likely to get within 3-10% of the trading range of the high/low of the day 1 out of 5 times. Not meaningful - which is why they're predominately only used by retail trader - so called "floor traders secrets" are never discovered by retail traders until most of the floor traders have moved on to something better.
 
You might as well generate random #s above and below the average - you're just as likely to get within 3-10% of the trading range of the high/low of the day 1 out of 5 times.
More...
Perhaps, but only you would know those numbers, and only you would be trading those numbers, and therefore unless you trade a tremendous size, you would be crushed the first time you started buying right below R1.

It is also important to remember that the numbers are pivot numbers, not highs, lows or closes. As fortuna posted, and as they are used, they represent support and resistance. The description in the link I posted is exactly what occurs. I was a local in the spoo pit. I still have friends who are there. Virtually everyone knows the numbers everyday.
 
Ummm... pivot...

OK... how did he test it...

Pivot Bar occurs then...

1. Next Bar is > or < then what point of the bar...

2. Next Bar broke Pivot Low or Pivot High's Extreme within X amount of barts...

3. ... 4. ... 5... etc. etc. etc.

One thing is how he tested the validity of the pivot.

He's gotta give the logic of the test then we can draw conclusions.

Gawd... just because on guy says he tested pivots and just the results, it doesn't give everyone else, at least, to draw conclusions of it.

USE YOUR HEADS! THINK!

Validity only exists only for the logic tested and there will be other validity on other versions of how the pivot was tested.
 
The excel file is quite simple. You can simulate the number of points by which u want the 5 pivot figures to be approached to give a valid top or bottom for the day.

I am currently back testing a model of intraday signals with these pivots to see what they bring in intraday. Model will be :

If Index is up, sell 1 point below first resistance, stop and reverse 7 points above this resistance, target profit 15 points etc...

I 'll be glad if s.o can share with me an excel visual basic code on such a simulation model to spare my time

Thanks
 

Attachments

you are missing an important statistic. here is some homework for all of you. find out how many days the pivot # (the middle one) was reached during the intraday session. now find out the average distance that price opened from a pivot each day. this should open some eyes. :) then see how FAR away price got on average BEFORE touching the pivot. even better ;)
 
Quote from funky:

you are missing an important statistic. here is some homework for all of you. find out how many days - or times ? -the pivot # (the middle one) was reached during the intraday session. now find out the average distance that price opened from a pivot each day. this should open some eyes. :) then see how FAR away price got on average BEFORE touching the pivot. even better ;)
More...

Hello Funky,

What do you suggest ? that prices tend to converge towards the pivot most of the time ?
 
You must log in or register to reply here.
Back
Top