What is the beset way to find the most recent date the market was open?

[ETFVoyageur]

Examples:

• I run it in the evening, and the market was open that day, I expect that date
• I run it during the day and the market is open then I expect that date
• I run it any time the market is not open I expect the last date the market was open

Is there any reasonably elegant way to do that?


Thanks,
EV

[NinjaTrader_ChelseaB]

Hello,

Thanks for your post.

The easist way to find the time of the last bar from an indicator is to use Time[0]. This will be the time of the last bar.

For example:
Print(Time[0]);

This would print the close time of the last bar in the Output Window.

[ETFVoyageur]

Hello,

Thanks for your post.

The easist way to find the time of the last bar from an indicator is to use Time[0]. This will be the time of the last bar.

For example:
Print(Time[0]);

This would print the close time of the last bar in the Output Window.

Thanks for the idea, but it is not quite what I need. II am trying to identify securities that did NOT trade the last trading day. Unless you have a better idea, what I need is to compare the time of the last bar to the last trading day, so I can flag mis-matches.

--EV

[NinjaTrader_ChelseaB]

Hi ETFVoyageur,

what I need is to compare the time of the last bar to the last trading day, so I can flag mis-matches.

--EV

That would work fine.

You may need to use the current bar's time and subtract a day (and use just the date not the time after this), and use the current time and subtract a day (and use just the date) for a good comparison.

[ETFVoyageur]

I'm still not with you. Suppose the date of the last daily bar is not today. How do I know whether the discrepancy is due to the market not being open, or to the security not trading?

--EV

[NinjaTrader_ChelseaB]

Hi ETFVoyageur,

You have the right idea of comparing the bar date to the date of the last session.

I am just thinking that you would need to do a little bit of work for that to work correctly.

I mentioned that you should use the current time and subtract a day. Really you will need to have the days that the instrument trades hard coded into the strategy to know if the previous day was a trading day.

[ETFVoyageur]

Hi ETFVoyageur,

You have the right idea of comparing the bar date to the date of the last session.

I am just thinking that you would need to do a little bit of work for that to work correctly.

I mentioned that you should use the current time and subtract a day. Really you will need to have the days that the instrument trades hard coded into the strategy to know if the previous day was a trading day.

And that is what I am asking for -- how to determine the last trading session, whether it is today or some other day. I do not want to hard code it, because things change. Furthermore, any hard coding I do will not know anything about market holidays. There really needs to be a supported way to know when the market was open. Getting it right is too complex to be hard-coded into an indicator.

I suppose I could load a data series for something I know traded, such as SPY or an index and look at its last date. Not very elegant, and pretty darned heavyweight if I do it in places like Market Analyzers with 1600-2500 rows.

--EV

[ETFVoyageur]

To bring some closure to this thread, here is what I have done. It is a hack, and not mathematically satisfying. However, it also seems to be adequate for now. I wrote a simple custom indicator, specifically for Market Analyzer use, that does the following:

• It checks whether the date/time of the last bar is within the last 7 days. I'm not interested in anything that has not traded in the last week anyway. In practice, this seems pretty good at getting rid of the chaff.
• If so, it returns the closing price
• If not, it returns -1

You can then filter on:

• <= 0 to hide stocks not actually trading
• Any price you like to hide stocks that are not in the price range you like to trade

As I said, it is a hack. It would be better to compare to the last open market date instead of 7 days, but I now of no sensible way to do that. (I consider something like using a SPY data series and looking at the last date to be too heavyweight for big Market Analyzers.)

As far as I can see, this meets my needs for now. Any better ideas are always welcome, so speak up if you have a better idea.

--EV