e-signal dodgy data?

[gcb01]

e-signal/metastock dodgy data?

I use Metastock v9 and e-signal and have been doing some system testing today. For a number of reasons I ran the same test (exactly the same) about 40 mins apart and got 246 out of 276 stock with different (in most cases very different) results. Given that the system used for both tests was identical (I've checked) and the data start and end times I used were the same the only thing I can put this down to is dodgy data.

Anyone else come across this?

 

[anmolway]

Have to agree gc, esignal data is a bit Cr*p, especially figures for volume can be so misleading.

 

[Spreadbetteur]

I was using eSignal for FX, then found that there were odd price movements, which did not show in other charting packages. On some days there would be a series of spikes all to identical price, which did not appear elsewhere.

eSignal agree that their FX feed (which comes from multiple data feeds) needs work on.

In the meantime I have moved elsewhere !

 

[Roberto]

Likewise. I totally agree with Spreadbetteur and have done the same thing for the same reasons.

 

[gcb01]

more cr*p

Have to agree gc, esignal data is a bit Cr*p, especially figures for volume can be so misleading.

I've done some more work and the found hordes of daft spikes that are not reflected in the daily data. I've raised this with e-signal and will report on the outcome.

 

[FTSE Beater]

Hi gcb01

May I ask. What data you looking at, i.e. UK shares, indicies ?

Thanks

 

[gcb01]

Data

Hi gcb01

May I ask. What data you looking at, i.e. UK shares, indicies ?

Thanks

Sorry, should have said US stocks.

 

[gcb01]

For info

Sorry, should have said US stocks.

For one stock I examined in detail (FAST), over a period of 3082 15 minutes bars, 1059 had significent spikes exceeding the daily data high/low over the period. The average spike size was 39.03 against the average Daily data closing price of 59.32.

Is it me or is this nonsense?

 

[JayF_eSignal]

For one stock I examined in detail (FAST), over a period of 3082 15 minutes bars, 1059 had significant spikes exceeding the daily data high/low over the period. The average spike size was 39.03 against the average Daily data closing price of 59.32.

Is it me or is this nonsense?

We spot checked a number of different bars on FAST, and were unable to find any disparities. I think what you may have been seeing is pre or post market data outside of the daily range. This can certainly happen as the Daily High/Low range almost always ignores the pre/post markets. If you limit the 15 bars to only be 9:30a - 4:00p Eastern Time, then I do not believe there should be any differences.

Gcb or to anyone else posting in the thread, if you have any examples of the data issues being discussed we'd certainly like to research them further!

 

[JayF_eSignal]

I was using eSignal for FX, then found that there were odd price movements, which did not show in other charting packages. On some days there would be a series of spikes all to identical price, which did not appear elsewhere.

eSignal agree that their FX feed (which comes from multiple data feeds) needs work on.

In the meantime I have moved elsewhere !

Do you have any examples of this along with a comparitive from another charting package?

A few things to note with eSignal's Forex data is that our data is a composite from over 200 contributing banks, and what you see on a chart is actually an accumulation of bid data from all of these banks. If even one of these banks sends out an incorrect bid price on a currency pair, then that pair's chart will have a spike. Also keep in mind, when comparing against other packages that many times other packages only include one (or perhaps a handful) of banks. This lessens the chance of "spikey" data, but also provides a limited picture in the overall Forex market.

Again, if anyone has any examples that we can look at further, we'd be more than happy to assist (either publically or privately.)

 

[gcb01]

checked

We spot checked a number of different bars on FAST, and were unable to find any disparities. I think what you may have been seeing is pre or post market data outside of the daily range. This can certainly happen as the Daily High/Low range almost always ignores the pre/post markets. If you limit the 15 bars to only be 9:30a - 4:00p Eastern Time, then I do not believe there should be any differences.

Gcb or to anyone else posting in the thread, if you have any examples of the data issues being discussed we'd certainly like to research them further!

I have checked and the bars are limited to 9:30 to 4:00 EST and the problems occur at many bars other than the first and last. Happy to send you the data in question.

yuletide felicitations

 

[JayF_eSignal]

I have checked and the bars are limited to 9:30 to 4:00 EST and the problems occur at many bars other than the first and last. Happy to send you the data in question.

yuletide felicitations

Please feel free to send this data to [email protected]

Thanks!
Jay

(Happy Festivus & Merry Christmas to all)

 

[JTrader]

JayF_eSignal -

Do you have any examples of this along with a comparitive from another charting package?

A few things to note with eSignal's Forex data is that our data is a composite from over 200 contributing banks, and what you see on a chart is actually an accumulation of bid data from all of these banks. If even one of these banks sends out an incorrect bid price on a currency pair, then that pair's chart will have a spike. Also keep in mind, when comparing against other packages that many times other packages only include one (or perhaps a handful) of banks. This lessens the chance of "spikey" data, but also provides a limited picture in the overall Forex market.

Again, if anyone has any examples that we can look at further, we'd be more than happy to assist (either publically or privately.)

Hi JayF_eSignal

While several people on these forums have criticised the GTIS indicative prices forex datafeed - for the spikes etc. it is of credit to esignal that you are prepared to answer questions and deal with such potential criticisms in an external public discussion forum. Keep up the good work.

Perhaps the best solution would be for esignal to sign a datafeed deal with EBS and maybe Reuters for the provision of forex data. Providing that either EBS or Reuters data is provided for each currency pair depending upon which platform most volume for each currency pairs trades. Is this a viable possibility for esignal? if not, why not?

Many thanks

jtrader.

 

[JayF_eSignal]

Thanks for the post, jtrader. This is unlikely due to GTIS (our Forex data provider) being a sister company to eSignal under the Interactive Data Corp. umbrella. Here's a KB Article we put together discussing our Forex service and outlining why we feel an institutional feed like GTIS is so valuable.