Some people tell me that I have “too much time on my hands” because I spend so much time “crunching numbers.” I tell them “That’s ridiculous, I don’t have any time on my hands because I am so busy crunching numbers” (That usually shuts them up. At least for a little while).
In any event it is true that I am something of an “indicator junkie.” And it is also true that sometimes I multiply two numbers together (or divide two numbers) just “because I can.” And every once in awhile something potentially valuable seems to surface (see Applying VIX(like) Indicators to Stocks).
But in the most recent case I am hoping that someone can help me here. In the not too distant past I came across an indicator referred to by the acronym of MVCI – or at least that’s what I recorded it as. I just can’t remember where I read about it. In fact I don’t even recall what MVCI stands for.
But I like what I’ve seen of it so far.
The MVCI Indicator Calculations
OK, what follows is a list of a fairly lengthy set of calculations. If you are not a “numbers geek” you might consider skipping down to the actual results.
A = Daily High
B = Daily Low
C = Daily Close
D = 200-day moving average of daily closing prices
E = Daily True High (Today’s high or yesterday’s low, whichever is greater)
F = Daily True Low (Today’s low or yesterday’s high, whichever is lower)
G = Daily True Range (E-F)
H = Average Daily Price (A + B) / 2
I = 15-day Average of H
J = 15-day Average of G
K = MVCI = (C-I) / (J * Square Root of 2))
L = Buy Signal Cutoff Value
M = # days to hold a long position
Figure 1 displays an Excel spreadsheet with these calculations.
Figure 1 – Excel spreadsheet for MVCI (whatever that means) Indicator
Hmmm, maybe I do have too much time on my hands. But I digress. In (slightly long-winded) English, the indicator is calculated by:
1) Subtracting the 15-day average of the average daily price (defined as( [high]+[low]/2)) from today’s closing price, and dividing that result by;
2) The 15-day average of the Average True Range times the square root of 2
Don’t ask me how someone came up with multiplying something by the square root of 2 because, well, heck I don’t even know what MVCI stands for, so how would I know?
The default value for variable L is -0.51. The default value for M is 22 days.
A “Buy Signal” occurs when:
1) The closing price for SPY is above its 200-day simple moving average AND;
2) The MVCI value for that day is -0.51 or less. So when a buy signal occurs the trader buys SPY and holds it for 22 trading days.
If MVCI drops below -0.51 during these 22 days then the 22 day holding period starts again at 22, i.e., positions can be held for longer than 22 days.
Figure 2 displays the fluctuations of the MVCI since 12/31/2012 with the -0.51 level highlighted in red. 
Figure 2 – The MVCI (whatever that means) indicator in action
So in a nutshell, anytime SPY is above its 200-day moving average each day the MVCI is below -0.51 starts a 22 day holding period.
Crazy , right?
Well, maybe we should consider the results first.
The Results
OK, so SPY started trading in 1997, to get a 200-daymoving average we can start testing in June of 1998. If we start with $1,000 and invest in SPY for 22 trading days following any day where MVCI drops to -0.51 or below, we get the equity curve that appears in Figure 3. No interest is earned while out of the market and buying and holding SPY is also plotted in red.
Figure 3 – Growth of $1,000 using MVCI (blue) versus Buying and Holding SPY (red); 1998-2015
For the record:
-$1,000 invested in SPY using MVCI in the manner described here grew to $4,417 (+342%)
-$1,000 invested in SPY on a buy-and-hold basis grew to $1,848 (+84.8%)
Maybe not so crazy, right?
Summary
For the record, please note that I do not actually use this indicator at the moment in any of my trading, nor am I recommending that you start using it either. I have this rule – well, OK in the immortal words of Bill Murray “it’s more like a guideline” – that I don’t use indicators when I don’t even know what they’re freaking called.
But the purpose of this blog is not to offer “advice” or to tell you what to do. Its purpose is simply to educate you and give you some things to think about that you might not otherwise.
Like for instance, “What the heck does MVCI mean?”
Jay Kaeppel
Jay,
I would like to test this. Would you define daily true high and daily true low for the purposes of this indicator? Thx.
E = Daily True High (Today’s high or yesterday’s low, whichever is greater)
F = Daily True Low (Today’s low or yesterday’s high, whichever is lower)
Interesting. I made an indicator for this in ProRealTime based on your info. Email me if you’d like the code.
Also a quick eyeball makes me wonder if a slightly different entry signal would work better (haven’t tested it). What about waiting for confirmation, i.e. the day that the indicator moves above the -.51 mark would be the signal instead?
Sorry to hog the comments, but I love indicators too. 🙂 Setting a dual threshold of time as well as the MVCI indicator hitting a certain level seems to improve results over the basic rules. Holding for 33 days OR selling if the MCVI went over 2.5 increased the yield substantially.
Using a high MVCI reading as a profit-taking method is an interesting idea that I’ll have to take a look at (Great, more numbers to crunch…..thanks alot! Har) Jay
To many parameters and rules (probably obtained by data mining). I would be very surprised if this “system” will work in the future.
You may be right. Still it is intriguing as a variation of the basic “overbought/oversold” type indicator. For what it’s worth the last year and half (or so) is “real time.” Jay
http://acronyms.thefreedictionary.com/MVCI
You may have written it down wrong.
Well that is certainly a possibility (especially give my penmanship….)
I found an MCVI function and indicator in my TradeStation library. Guess I looked at it some time in the past. It is the Modified Chartmill Value Indicator (MCVI). If you want I can post the TradeStation code. I have no idea where I got it (probably the TradeStation forum, or possibly another blog).
To clarify, the MCVI is the same as the indicator you posted. It was in the May 2013 issue of Tech Analysis of Stocks & Commodities magazine.
The mystery is solved! Thank you Gary. I knew I read it somewhere. So is the May 2013 issue came out in April sometime we can say that since say May 1 of 2013 is “real time”. Jay
Jay, do you have a contact form somewhere that I can email you? I have some technical questions about this that I don’t want to clog the comment feed up with. Or perhaps you could email me using the email address I’ve supplied in the wordpress field? Thanks.
If you want to boost your cred a little bit and defend against the data mining charge then one thing you could do is study the performance across a range and report total profit or some other value. That is, don’t just show us the results for -0.51. Show us the results for -0.70 to -0.30 or something. I like what you did here but without having any idea what the surrounding parameter space looks like, I have no clue whether this was total randomness or the “well-selected example” or otherwise.
Mark, points well taken. For the record, the -0.51 value was the value used in the original Stocks & Commodities article from which I got this indicator. Jay
Test it out on other markets to see if it has any real value or is just the product of data mining. That, and test out various parameters to see if it’s been overly optimized. It looks curve-fit to me whenever I see stuff like -0.51, 22 days and 15 days. Those numbers don’t make intuitive sense (5, 10, 20 days make sense as it’s 1 week, 2 weeks, 1 month of trading etc).
That is actually very good advice. The parameter values used in the article came directly from the Stocks and Commodities article where I first read about this indicator. Whether reading my stuff or someone else’s it is always a good idea to “do some homework” before risking real money. Jay
Well, using MarketXLS works for me. It’s great