Difference in indicator value between fidelity & yahoo data.
Author: bobydesi123
Creation Date: 2/13/2010 11:39 AM
profile picture

bobydesi123

#1
Hi,
I am having one of those days! Sorry for asking trivial questions. I am looking at s&p500 index, .spx from fidelity and ^GSPC from yahoo, daily scale. I have plotted 200ema.Modern on both of them. On 2/6/10 o/h/l/c is same for both yahoo and fidelity. I have randomly compared the o/h/l/c for .spx and ^gspc and they agree. However, on 2/6/10 the yahoo chart clearly shows support @200ema=1047, where as the fidelity does not. Fidelity has value of 1037 for ema200! Obviously, there seems to be difference between these two datasets. How do i figure this out. I was looking at an article that has a chart for sp500 that also seem to agree with Yahoo, however, i dont know their source of data, it could be some other datasource or yahoo. I am totaly lost. Time for more cofee!Thanks for ur help in advance.
profile picture

Eugene

#2
Obviously, there can be differences in data by different data vendors. It happens.
profile picture

bobydesi123

#3
Eugene,
I understand that, but in this case the difference in data makes a big difference(pun?)! As far as the market was concerened on 2/6/10, around @.SPX 1047(yahoo's ema) lots of buyers came in and the market turned up, as if quite a bit of traders beleived that 1046-1047 was the SUPPORT!
However i know that u have nothing to do with fidelity data. I also wanted to know if people here noticed similar behaviour or my data is corrupted. Thanks for ur help.
profile picture

Eugene

#4
Hmm. Firstly, 02/06/2010 was a holiday - so you must be talking about the 02/05/2010 bar, right? There was a touch of support (represented with the 200-day EMA) on both Y! and Fidelity charts.

Try right-clicking on the chart and selecting "Reload chart history".
profile picture

bobydesi123

#5
Eugene,
u r right about the date, my mistake, it is 2/5/2010. I "reloade" as u had suggested and worked fine. This bring up another question, I had disabled "update data on demand". However, update EOD data everyday. Should not the updating of data correct bad data?
profile picture

Eugene

#6
Yes, Yahoo data is auto-corrected when it's changed on the server, but that usually affects Volume. When WL5 performs an update, the provider requests the last 5 days of data and checks for corrections. (So, 02/05/2010 falls out of the range.) Also, corrections are applied when you have "Always return data with partial bar" enabled.
profile picture

bobydesi123

#7
Got it. I cant say this often enough that ur support is excellent.
profile picture

cdkoss

#8
Stockcharts.com just had a great article last week on how ema's are calculated..actually very complicated. It is based on the amount of past data the vendor uses...many don't use full data base to save on space so don't get exact number. Stockcharts claims they use the full data history of a stock so theirs is the exact ema.
profile picture

Cone

#9
EMA isn't a particularly difficult indicator (the calculation is only about 3 lines of code), but as I just stated in another thread, EMA is part of a group of indicators that are "unstable", and therefore the exact values depend on the starting point of the data. For a full discussion and demo, see the WealthScript Programming Guide: Indicators > Stability of Indicators