Zool wrote:More on pivots: I've checked the pivots on USDCHF. While using Bob's pivot indicators the order are this (low to high): monthly - weekly - daily (showing clear uptrend)
Using other pivot indicators the order: monthly - daily - weekly (ranging?)

All of this I will write about is made on Cowboy IBFX (where the server time zone is gmt+3 Bob's pivot indicators show the same value as other ones). So I've checked the weekly pivot value. Manually calculating it based on the weekly candle it should be correct. The weekly candle on USDCHF shows the following values:
H: 0.96949
L: 0.94791
C: 0.94975

Calculating the pivot values with (H+L+C)/3 formula gives 0.95572. This is the value Bob's weekly pivot indicator shows. The glitch is that on Cowboy IBFX the weekly candle's closing value is not the friday close value, but the sunday close value. I don't know why
When Cowboy IBFX starts a new week with it's GMT time zone there is already 09:00 in Sydney, the market started 2 hours ago, how could be the end of sunday used as a weekly close value? It was not really apparent till now, because there was no/small gaps in the weekend.

When not using the sunday candle to determine the weekly pivot the values are:
H: 0.96949
L: 0.95126
C: 0.96062
Look at the big difference in the low and close values. The pivot based on these numbers is 0.95046 and this is the same as other indicators and online pivot calculator sites showing. I think this is the correct value, because a trading week starts at Sydney open and ends with NY close, if using Cowboy IBFX's weekly candles that's not the case. The weekly pivot would be off for the entire week, and if there's a weekend gap again, then the next week will be wrong too. Maybe the difference is not that big in the majority of times, but this week they are noticeable.

There are daily and weekly pivot indicators made by Cowboy IBFX and those are showing the correct values. If someone, who can code may look into those indicators source, then maybe the fix can be made to Bob's indicators ( if needed).

nanningbob wrote:
Now you know why I called it a test version I guess the question really becomes on how we handle weekend gaps. Do we include them or not include them?? That seems to become the real issue. Without the gaps the differences are small. With the gaps they can be quite different.

spotdespot wrote:

nanningbob wrote:
Now you know why I called it a test version I guess the question really becomes on how we handle weekend gaps. Do we include them or not include them?? That seems to become the real issue. Without the gaps the differences are small. With the gaps they can be quite different.

I think any fix that took in to account the Friday close value (rather than Sunday close) as Zool suggests would solve any debate wouldn't it? I don't care about 10-20 pips too much but 60 pips makes a significant difference imo. I know this latest gap was a little extreme but as pivots are somewhat self-fulfilling in terms of use as S/R it would seem sensible to be using pretty much the same pivots as everyone else - as long as the "fix" isn't too onerous.

nanningbob wrote: Actually none of this discussion is new. We had the same issues with the weekly open histograms all for the very same reasons. Different start times for criminals, some have a Sunday candle some dont, do you include the pre gap or post gap open, etc etc. I came to the conclusion that I cant design my system to fit all these variations. Each trader would have to do their own work around. What ever solution I propose would just cause a different problem for someone else. So, I decided I had to use what worked for me, the criminal I used, and the markets and time zones I traded. NZ and AU people have different schedules than those from other parts of Asia, who are different from those from Europe and Africa, then the Western hemisphere people. Each trader is simply going to have to adapt. And in reality this would be true for all indicators not just the ones in this system.

nanningbob wrote:
I tried eyeballing your chart with mine. Except that mine updated for the Monday gap the rest seem about the same. In other words your Tues.....

spotdespot wrote:Hi Bob - garyfritz actually came up with a solution to the weekly open histogram issue - which works with any timeframe criminal (GMT or otherwise) -

garyfritz wrote:They just get the previous week/month candle using iShift, then pull out the H/L/C with iHigh, iLow, iClose. I don't see any fix there.

garyfritz wrote: Or maybe it doesn't really matter. As someone (Spot?) said, there are a zillion different pivot formulas. They can't all be right, or at least any more right than almost any other. Maybe how you trade around the pivot is more important than where the pivot is.

garyfritz wrote:Or maybe it doesn't really matter. As someone (Spot?) said, there are a zillion different pivot formulas. They can't all be right, or at least any more right than almost any other. Maybe how you trade around the pivot is more important than where the pivot is.