My 21 day moving average isn't right

[patpat]

I have a friend who uses NinjaTrader, and we have identical settings. We use the same number tick chart, the same 21 day moving average. But sometimes price breaks above the 21MA on my chart, while it bounces off the underside of the 21 on his chart. I understand that candles can vary, but I am surprised that MA do not act the same.

[NinjaTrader_Kyle]

Hello,

Thank you for your post.

I would recommend that you compare the OHLC values in the chart with the bars in the other parties chart. As you mentioned, there can be differences in tick values between providers - since the tick data is used as the input for your simple moving averages differences in bars can lead to differences in indicators.

[koganam]

I have a friend who uses NinjaTrader, and we have identical settings. We use the same number tick chart, the same 21 day moving average. He took a nice trade today when price pulled back to the 21 MA. But on my chart, price pulled back to the 21 MA, but then exceeded it by quite a distance, relatively speaking. The only way I could simulate his set-up was to include a 75 day MA on my chart. (i.e., price on my chart pulled back at the 75 day MA). As you can see, his 21 day MA acted quite differently from mine.

I realize that candles may not form the same, but I do not understand why the 21 MA would not act the same.

Is there a setting on my charts that I can adjust? Do you think this is normal?

The EMA is an infinite impulse response filter. If your signals depend on any kind of infinite impulse response filter, your charts MUST start on exactly the same bar, or you are liable to get different results. Check that you are both starting on exactly the same bar: not just the same date.

[patpat]

I forgot to mention that we have the same data provider, as well. Thanks for your time. The situation I described is probably normal.

[patpat]

thanks. My trading does not rely on MA signals.

[Harry]

I have a friend who uses NinjaTrader, and we have identical settings. We use the same number tick chart, the same 21 day moving average. He took a nice trade today when price pulled back to the 21 MA. But on my chart, price pulled back to the 21 MA, but then exceeded it by quite a distance, relatively speaking. The only way I could simulate his set-up was to include a 75 day MA on my chart. (i.e., price on my chart pulled back at the 75 day MA). As you can see, his 21 day MA acted quite differently from mine.

I realize that candles may not form the same, but I do not understand why the 21 MA would not act the same.

Is there a setting on my charts that I can adjust? Do you think this is normal?

There are several possible reasons for this behaviour:


Different data feeds

If you talk about instruments that are not traded on a central exchange (futures), but look at FOREX, then the data which is used as input for your chart can be different. Another possibility are slightly different time stamps. If you use broker data which comes without a time-stamp, then data will only be stamped by your PC, which will produce a different time-stamp compared to the one which was originally attached to the tick by the exchange.


Different trading sessions

Intraday charts: NinjaTrader cuts off the current bar at the end of the session and/or the trading day leaving you with one incomplete bar. This may not be the case for your friend's data feed. Different input data will result in different closes and have an impact on the indicator values.

Daily charts: For futures daily charts may use the settlement price (e.g. NinjaTrader + Kinetick) , the regular close (e.g. NinjaTrader + CQG) or the full session close (e.g. Interactive Brokes), so you need to find out what is displayed on your chart. For FOREX some datafeeds use Cyprus time, others use GMT or EST to determine their daily charts. Obviously this leads to different closes.


Different lookback periods

As @koganam stated correctly, the EMA is an IIR filter, because it relies on a recursive formula. While your 21-period SMA will be accurate after 21 bars, the 21-EMA may require a training period of several hundred bars, before it becomes accurate.


Explaining the difference

If you wish to understand the differences please

-> compare all the bar closes that go into the calculations
-> check the datafeed which is used
-> check the session templates used
-> make sure that you both have a sufficient number of bars on your chart

[koganam]

thanks. My trading does not rely on MA signals.

Not to cavil, but your initial query was talking about a difference in appearance of the 21-day moving average. I thought that that was what you were asking about. My apologies for not understanding what you were really talking about.

[patpat]

dang! I hate when I leave out pertinent information! I trade the ES.

[Harry]

dang! I hate when I leave out pertinent information! I trade the ES.

And then tell us,

- whether you used a 5 tick or a 1,500 tick chart
- which session template you had applied
- who is your data supplier
- whether you observed it on historical or real-time data
- what was the lookback period of your chart

and best attach the two charts (yours and the chart of your friend).

[patpat]

Thanks for your replies. I have figured out why my chart didn't look like my friends. Unbeknownst to me, my chart changed from a 1500 tick to a 1500 volume. I noticed other changes on my chart, as well, and couldn't figure out what happened. So the mystery has been solved. Thanks again.