Higher/lower timescales not lining up in charts or backtests  [SOLVED]

I'm attempt to both 15-minute and 1 day bars for CL in a strategy I'm testing. It appears that the daily bars have a timestamp of 22:59, and so are lining up with the 15-minute bars at 22:59. However the CL electronic market closes at 17:15, and the daily bars should line up with this time instead. I attempted to adjust the session hours in the Instrument properties, but this had no effect.

I've printed output from the calcBar() in the method from my strategy which contains the following output method calls. A snippet from the output. As can be seen, the daily bar doesn't change until midnight,even though the close price on the daily bar matches the close price of the 17:15 15-minute bar.

Any help would be appreciated.

Thanks,
-Rob

I attempted to adjust the session hours in the Instrument properties, but this had no effect.

Hello Rob,

The bars were already written to the database with the session setting specified at the moment the data was requested from the data provider. After changing the session settings all historical bars remain the same, all new bars are affected by the changes. In order to apply the new session settings for historical bars you need to plot them on the chart and reload the chart.

Hi Henry,

so I refreshed the chart, and the time on the bars is now correct, however the price is not lining up between the 15-minute and 1 day bars. I would expect to see the close price of the last 15-minute bar of the day to match the close price of the daily bar, but this is not the case. I using symbol QCL# (IQFeed), and I have set the Session time to CME Commodity Futures (Globex), I've tried NYMEX Energy (Electronic), and default as well. With the Globex session I see the following differences between the close of the last 15-minute bar of the day and the daily bar close.

3/14/13 - Difference $0.05
3/13/13 - Difference $0.25
3/12/13 - Difference $0.04
3/11/13 - Difference $0.18

Thanks,
-Rob

Hello Rob,

The data is provided by IQFeed "as is". The difference between daily bar and intra-day data is expected You can come to our live chat and we will help you to check the historical data provided by IQfeed.

You can get a daily bar with customized sessions if you create N minute chart with your own session template. IQfeed also provides ability to build minute and daily bars out of tick data:
Quote manager->Tools->Data sources->IQFeed->Settings->"build minute and daily bars from tick data". To make this setting come into effect you need to restart MultiCharts and all of its applications.

Please keep in mind that IQFeed has a limitation for amount of provided historical data:
Tick data- 8 calendar days within session hours, 120 days in non-session hours,
Minute data– up to 6 years (@ES# ~4.5 year)- only TRADE data,
Daily data – since 1901 (@ES# ~13 years), (INDU.X since 1928 year)

That means if you select building all resolutions from ticks then you will have less historical data, compared to using minute and day resolutions.

Another case of this issue.

Please vote here on this.

https://www.multicharts.com/pm/viewissu ... _no=MC-856

Hello Rob,

The data is provided by IQFeed "as is". The difference between daily bar and intra-day data is expected You can come to our live chat and we will help you to check the historical data provided by IQfeed.

You can get a daily bar with customized sessions if you create N minute chart with your own session template. IQfeed also provides ability to build minute and daily bars out of tick data:
Quote manager->Tools->Data sources->IQFeed->Settings->"build minute and daily bars from tick data". To make this setting come into effect you need to restart MultiCharts and all of its applications.

Please keep in mind that IQFeed has a limitation for amount of provided historical data:
Tick data- 8 calendar days within session hours, 120 days in non-session hours,
Minute data– up to 6 years (@ES# ~4.5 year)- only TRADE data,
Daily data – since 1901 (@ES# ~13 years), (INDU.X since 1928 year)

That means if you select building all resolutions from ticks then you will have less historical data, compared to using minute and day resolutions.

Build daily + Minute bars from tick data is not a suitable solution for me. It is a limitation of the data provider since only 120 days are available. There should be an option "Build Daily Bars from Minute Data". Those daily bars that are built should be associated with a session since there are obviously different ways to build a daily bar depending on what session was in use when the bars were loaded.

Agreed that building by tick isn't a viable alternative due to the 120 day limitation. As far as I can tell, the close on the daily bars are accurate as I have compared the IQFeed data to CSI data and also Interactive Brokers, so I'm not sure why such a huge difference in the 15-minute bars (and even 30-minute) bars. I'll take a look at the 30-minute bar historical data from Interactive Brokers and compare to IQFeed and see what I find.

From what I see though, I don't think I would want to build daily bars from 15 or 30 minute bars due to the inaccuracy of the intraday bar close prices.

Agreed that building by tick isn't a viable alternative due to the 120 day limitation. As far as I can tell, the close on the daily bars are accurate as I have compared the IQFeed data to CSI data and also Interactive Brokers, so I'm not sure why such a huge difference in the 15-minute bars (and even 30-minute) bars. I'll take a look at the 30-minute bar historical data from Interactive Brokers and compare to IQFeed and see what I find.

From what I see though, I don't think I would want to build daily bars from 15 or 30 minute bars due to the inaccuracy of the intraday bar close prices.

The reason for the non matching of the daily bars to the intraday bar closes is likely because the daily bar uses the settlement price that has been released by the exchange after close while the intraday bar uses the actual close.

In my opinion the intraday bar is more accurate. I've never used settlement prices.

Good point Dave, it makes sense that they are using the settlement price provided by the exchange. Since the close price for the daily bar isn't really accurate then, the daily bar's Open price should still match the open price of the first 15-minute bar of the session, which I hadn't checked up to this point, and it is in fact pretty close (within 0-2 cents for CL).