Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
Mike, just spent the last hour or so on McClellan. The first thing I realize is that I need to ensure some type of data accuracy. The problem is that on a daily basis, there is a single data point that matters for a/d--the closing value. And when I compare the "official" numbers from McClellan Financial Publications, from 9/28 to 3/22, from their spreadsheet, I get a sum of Oscillator values of 310. When I use the barchart data from Sierra and sum the values over the same period, I get a value of about 1000. Not too bad, but a little further off from where I'd like it to be. By the way, it took me a minute to realize this because my numbers were way off, but you need 40 more bars on the chart when using mcclellan due to its using a 39 and 19 period EMA, so if you want the right oscillator value on Jan 1, chart should start Nov 1 or so.
A comparison of the oscillator values to a known source:
These should match up in absolute value and shape, and they appear to be pretty close.
Now, the summation index. The absolute values are off due to the difference in start date from which the values are summed, but the shape is similar, though you can see some obvious differences:
So, McClellan Oscillator takes the 19 period EMA of the A-D and subtracts the 39 period EMA of the A-D. That's it, plain and simple. There is a ratio-adjusted method which takes a ratio based on the number of NYSE stocks; this is for long term comparison because a pure A-D does not take into account the fact that the number of listed stocks changes over time, so theoretically the ratio adjusted version could be used for very long comparison, say, several decades I would guess.
The Summation Index is simply a running sum of the Oscillator values.
The idea behind advancers and decliners as a measure of breadth is that if the index is moving up, it can continue to do so with large caps doing most of the pulling. However, this cannot continue for an extended period of time, as eventually the small caps must contribute to indicate overall "health." So, a declining summation with an advancing index may indicate that a correction is coming, because the entire market is not helping as much as just the bigs.
Just as is the case with all divergence indicators, the question is, how long as you willing to go against the market while it diverges? It can diverge many times, for many months or even years. And intraday, I have not used or looked so I'm not sure. But I might be willing to look at this as a long term investor, but on a daily basis as a day trader or even a swing trader, it's just too large of a perspective. The calculation itself lags 19 and 39 days (though it is an EMA and not an SMA), so I would look at this on a daily basis as an overall health reading, and only use it if I were going to be looking to take a trade that might last several months. But intraday, I will have to give a look and see if I see any edge there.
Though the recent data as shown above is fairly similar, look how different the values are from early 2011 to present:
Mike, perhaps you or someone could post your McClellan Summation index from January 2011 to present (as you know Mike, the study name in SC is "McClellan Sum. Index - 1 Chart")
I have no idea as to the accuracy of my intraday A-D data, and again, since these intraday values, like TICK, are dependent on data provider, I'm a little wary of putting too much trust in them. It's the same reason I'm wary of OTC forex, just too much room for one-offs and no centralized verified values. I am amazed at how many traders here, some of whom are very good, post charts with delta values and the like (like in footprint charts, cumulative delta, etc) that are just plain wrong, and yet they use the data and think it's helping them, when in fact it's more like a sugar pill used to cure a disease--effects are only in the mind So, forgive me for this long post but I want to ensure that I'm not using data that is too far from some reasonable median value.
What time frame do you want from January 2011? Daily?
Regarding intraday usefulness, since we're using looking at intraday adv/decl I would say it has the same usefulness as any other intraday analysis would based on breadth. You would of course need to use the intraday reading, not the daily reading.
I suppose one could make their own custom oscillator using a MACD or perhaps something else, to be able to lower or customize the values if you wanted.
Thank you Mike. As I suspected, my historical AD data is quite off. In January 2012, it should just shoot up dramatically as yours does, but mine stays flat. However, the more recent data seems to be quite similar, so at least that's good news. Okay, I'm going to step back from McClellan for a second and go to the source, just AD. Here is a 5 minute ES on top, A/D on bottom. Data is missing from 3/21 and I can't do anything about that but otherwise it's all there. How similar is our data?
The most obvious divergence is today, 3/22, with A/D trending down most of the day while ES kept pushing up. I will look at some more days and time frames next.
This will depend on the first start date. The first value for the summation will be the first value for the oscillator, and so the first date on your chart will be close to zero, and from there it just sums, so it might have been very negative at first for example. Also, part of the deal is that declining issues could outnumber advancing yet the market still gets propped up for quite some time.
So, forgive me for this long post but I want to ensure that I'm not using data that is too far from some reasonable median value.
