MFI Study in Q6.0

[svrz]

Hello

It seems that the MFI study in Q6.0 is significantly different than the MFI study in Q5.1. A few years ago I asked technical support how this indicator was calculated. The answer was the following;

Typical Price = (high + low + close)/3

MFI = RSI(TYpical PRice, n) where n = period.

Is the new MFI calculated differently than the above formula? If so, please elaborate.

Thanks in advance.

[ACM]

svrz
As far as I know the Money Flow Index is calculated as follows

Typical Price = (High+Low+Close)/3
Money Flow = Typical Price * Volume
If today's Typical Price is greater than yesterday's then Money Flow is considered positive.
If today's Typical Price is less than yesterday's then Money Flow is considered negative.
The positive and negative Money Flow are then individually summed for the selected period of time
Money Ratio = Sum of positive Money Flow / Sum of negative Money Flow
Money Flow Index = 100-(100/(1+Money Ratio))

I believe QCharts 6.0 is using this method for the Money Flow Index.
The following screenshot shows the Money Flow Index plotted in in QCharts 6.0 and the result of a formula I wrote in eSignal based on the logic described above. As you can see the results are the same
Hope this helps
Alex

Originally posted by svrz
Hello

It seems that the MFI study in Q6.0 is significantly different than the MFI study in Q5.1. A few years ago I asked technical support how this indicator was calculated. The answer was the following;

Typical Price = (high + low + close)/3

MFI = RSI(TYpical PRice, n) where n = period.

Is the new MFI calculated differently than the above formula? If so, please elaborate.

Thanks in advance.

[svrz]

Alexis

Thank you for the reply and describing in detail the implementation of MFI in QCharts 6.0. However I was describing the differences in the MFI study between QCharts 5.1 and 6.0, not any of the ESignal products.

In any regard, attached are screen shots of the study in version 5.1 and 6.0 that highlight the differences.

If we take a few sample readings of the data, this is what we see:

Instrument:YM07Z Interval: 5 minutes

Time QCharts 5.1 QCharts 6.0
-------------------------------------------------
11:10 47.77 68.12
13:00 45.66 70.84
14:45 55.13 39.04

As we can see from the data, the implementations of the MFI indicator in QCharts 5.1 and QCharts 6.0 are vastly different.

Can you kindly ask Mr. Condra to provide some feedback?

Thanks very much in advance.

[svrz]

QC 6.0 screen shot.

[ACM]

svrz
To clarify I was only replying to your question as to whether the MFI was calculated differently from your formula and if so how. I used an eSignal chart solely to confirm that the method to calculate the MFI that I described was indeed the one implemented in QCharts 6.0.
Alex

[JayF]

If I can step in....during the development and subsequent testing of 6.0, a bug was found in the way the money flow index was returning values. At times it was spiking dramatically, and so we looked into the calculation of this study. In the process of doing so, we corrected the spiking problem and also found an error in the calculation. That could be the reason for the calculation difference between 6.0 and 5.1.

In essence, QCharts is now matching the commonly accepted method for calculating Money Flow, which Alex was kind enough to post.

[svrz]

Alexis, thank you for the explanation.

Jay, thanks for clearing up this issue.

Regards

[svrz]

Jay

I apologize for reviving an ancient thread.

However, would it be possible to revert the MFI back to they way it was in version 5.x? The previous version normalized the price/volume magnitude which made the recognition of important support and resistance levels quite simple. Unfortunately the MFI indicator in 6.x fails to do that.

I fully understand that more important and pressing issues currently demand your staff's attention. However if this issue can be re-visited after the release of 6.1, I would greatly appreciate it.

Thanks in advance.