11 replies to this topic

[inamosa]

...but perhaps this is the most interesting one:

ADT = Advancers / (Advancers + Decliners)
ADT5 = 5-day average of ADT

Top 20 ADT5s Since 1970

(historical data provided by an Alabama-based firm)

I think everyone should note that breadth thrusts are not always successful on 1- and 3-month time frames (although they usually are), but breadth thrusts are almost always successful on 6- and 12-month time frames. The above one is so powerful that performance is also stellar on a 24-month time frame. We have just completed one quarter since the last Zweig Breadth Thrust now, and the performance to this point has been +0.7%. This bodes well for this quarter.

I find it interesting that the Nasdaq Composite MCOs are almost as high as they were in March 2009 (and July and September 2010) - which also represent levels that are very close to the highest in over 30 years.

If the market reflected the fundamentals as perceived by the public too much of the time, how would the pros make money? All that matters on multi-quarter and shorter time frames are what the large players' time and liquidity preferences and how one can determine said preferences.

"It is utterly useless for us on the outside, who buy and sell comparatively small blocks of stock, to conjecture about what 'they' are doing. We cannot know what the insiders intend to do, but we can see their orders on the tape when they execute them. That is why my plea is for everyone of us to have no mere opinions of his own, but to allow the actions of the market to tell him what is passing."

-Humphrey B. Neill in Tape Reading and Market Tactics

Edited by alysomji, 01 July 2011 - 07:54 PM.

[IYB]

Amazing numbers ALY. Strongest 5-day thrust ever. Pretty darned amazing. I touched on this subject today, too, at SS.com, but have been busy with a family reunion and hadn't had time yet to dig in deep into the current "breadth thrust" data. Your timely input on this is very highly appreciated. Really valuable stuff. If you have other numbers, tables, comparisons, I'd love to see them. And if I can find the time this weekend I'll add some thoughts on this to yours, fwiw. Thanks very much for this post and data. Regards, D ------------------------------ P.S. Your Humphrey Neill quote sounds just like what I keep iterating nearly every single day.

Edited by IYB, 01 July 2011 - 08:29 PM.

[nimblebear]

It's on super super low volume. Meaningless, as the market will soon tank beyond everyones belief.

[Slothrop]

Here's what I'd like to see: how the market has performed after mentions of the Zweig Breadth Thrust on Traders Talk.

[arbman]

Aly, the breadth data from past is useful, but I do not think it compares exactly to the ones today. Due to the computerized trading, the order flow is very even now and they can push all stocks together without an amazing amount of liquidity. Although the liquidity in this market place is pretty strong. Given that the breadth measures the broad flow of liquidity, it is obvious how the computerized trading of today cannot possibly compare to the data from past. For example, if you can find the data, find the program trading volume on NYSE over the past 10 years and see how the breadth thrusts behave in proportion to the program trading. For one there was less volume today than any day of the week [edit: for the points gained], and we had a record reading on NYSE. I am not saying the market will not go higher or the market is weak, I am only saying from a quantitative perspective, the past stats do not seem to be a good comp to me going forward. Probably the data from 2004 and on is a good comparison. Why? If you get a chance, examine the VOLUME of tick data from 2004 until today and you will see, the amount of tick activity has grown exponentially, the exchanges are trading almost all instruments now although the volume has not proportionally grown the same way. This tells me that, once again, the computerized trading is evenly distributing the liquidity very very efficiently today than say merely 5 years ago, and it is exponentially better than say 10 years ago... All in all, the markets are actually much much stronger due to the better order management and should be less prone to manipulation, however the biggest manipulation is happening at the liquidity level from Fed, so I have to go with the flow. This is why it seems to me the cyclical studies of the breadth movements appear to be a more valuable approach than just the daily breadth readings during the strong initiations. This is what generally do and I see a bullish picture too...

Edited by arbman, 01 July 2011 - 10:52 PM.

[inamosa]

Aly, the breadth data from past is useful, but I do not think it compares exactly to the ones today. Due to the computerized trading, the order flow is very even now and they can push all stocks together without an amazing amount of liquidity. Although the liquidity in this market place is pretty strong. Given that the breadth measures the broad flow of liquidity, it is obvious how the computerized trading of today cannot possibly compare to the data from past.

For example, if you can find the data, find the program trading volume on NYSE over the past 10 years and see how the breadth thrusts behave in proportion to the program trading. For one there was less volume today than any day of the week [edit: for the points gained], and we had a record reading on NYSE. I am not saying the market will not go higher or the market is weak, I am only saying from a quantitative perspective, the past stats do not seem to be a good comp to me going forward.

Probably the data from 2004 and on is a good comparison. Why? If you get a chance, examine the VOLUME of tick data from 2004 until today and you will see, the amount of tick activity has grown exponentially, the exchanges are trading almost all instruments now although the volume has not proportionally grown the same way. This tells me that, once again, the computerized trading is evenly distributing the liquidity very very efficiently today than say merely 5 years ago, and it is exponentially better than say 10 years ago...

There are some problems with your argument that "this time is different," arbman. Here are a few of them:

Due to the computerized trading, the order flow is very even now and they can push all stocks together without an amazing amount of liquidity. Although the liquidity in this market place is pretty strong. Given that the breadth measures the broad flow of liquidity, it is obvious how the computerized trading of today cannot possibly compare to the data from past.

1. You're contradicting yourself here. On the one hand, you say the market can be pushed at once. However, this should not be possible in a very liquid market - which you concede this is. So, which is it? I think we need to be clear on the fact that high-frequency traders and other liquidity providers are normally delta neutral by the time the market closes each day - and that is also when breadth numbers are compiled. The exception to this is that some liquidity providers will take an active position overnight or longer if they have an extremely strong edge based off of what professional buy-side money is doing - however, this has been the case throughout market history.

2. Between the beginning of 1988 and the end of 2007, there was, for the most part, a distinct absence of breadth thrusts on the SPX despite a massive increase in computerized trading. This includes your 2004 period.

3. Recent breadth thrusts have been performing as well as they were prior to their long absence during the '90s and for most of the '00s. This is my biggest problem with your argument, and the only one that really matters. The proof is in the pudding. So far, the performance numbers do not back up what you have to say.

By the way, keep in mind that an extreme move in breadth based on the last 40+ years of historical data will also qualify as an extreme move (more extreme, in fact) on historical data since 2004 or whatever year you'd like to choose. We are talking about breadth changes that are more than two standard deviations from the mean on data going back to 1970. They are extremely rare in the big picture, although, not unlike today, there were clusters of them in the late '70s and early '80s as well (near the end of the last secular bear market) - interestingly enough.

Also, I have yet to see a useful or meaningful relationship between volume during the occurrence of a breadth thrust and the success of that breadth thrust when looking at backtests. Perhaps the thresholds for meeting breadth thrust requirements are so high so as to not cause an issue in this regard. Or perhaps its because the shortest time frame in which a breadth thrust can occur is two days, and that multiple days of low volume market activity are still sufficiently meaningful. Whatever may be the case, there is simply lack of evidence of a meaningful relationship.

Let's talk more facts and less opinions, folks - or at least back up your opinions with facts.

[orange]

The above one is so powerful that performance is also stellar on a 24-month time frame

Hi Alysomji,

Were you implying there is a relationship between the ADT5 and the 12/24 month gain (or was it just a statement)? I don't see a relationship.

Do you know of any methods to rank the strength (ADT5) that might predict the likely percent rise over a 12 month period?

I've never studied Zweig thrusts in depth to see if there are indicators that may give clues to the expected 12 month high. I wonder if there is a relationship between the 5day ADT and how oversold the market was 5 days before. I'm willing to speculate there is a relationship between the ADT5 and how oversold general indicators are 5 days before.

Edited by orange, 02 July 2011 - 02:27 AM.

[inamosa]

Were you implying there is a relationship between the ADT5 and the 12/24 month gain (or was it just a statement)? I don't see a relationship.

Very few types of breadth thrusts have led to performance, like shown above for the highest twenty ADT5 readings, 24 months after occurrence. It's an observation.

Do you know of any methods to rank the strength (ADT5) that might predict the likely percent rise over a 12 month period?

I've never studied Zweig thrusts in depth to see if there are indicators that may give clues to the expected 12 month high.

I wonder if there is a relationship between the 5day ADT and how oversold the market was 5 days before. I'm willing to speculate there is a relationship between the ADT5 and how oversold general indicators are 5 days before.

Of course, most breadth thrusts come after the market has been severely oversold. This is why breadth thrusts normally occur either during / after pivotal bull market corrections or at the beginning of bull markets.

A breadth reading is only one variable and I don't think you will be able to accurately predict the SPX level, say, a year from now, using it alone. There are a lot of other factors that I'd suggest come into play, such as monetary conditions, sentiment, etc.

[arbman]

First of all, thank you for sharing your analysis and I am not trying to discredit your work or anything, if it appeared that way... I agree that my objection doesn't stick without the data backing them up, but I do have the data. I think I need to clarify, all I said was the program trading allows more and more issues to move just a bit with the overall market and hence the recent thrusts look relatively the strongest among your historical comps. I just don't think this is the historically strongest thrust in terms of its trending properties, but this may be the best organized one to get the most bang for the buck and I think the thrusts from the 2009 lows were stronger than this one. I cannot disclose my methods in measuring the thrust and exhaustion patterns though and I still think this is a strong thrust and market too. We may have a year to top the bull market too like you said, I have a strong(er) rally from the end of September...

Edited by arbman, 02 July 2011 - 02:56 AM.

[inamosa]

I just don't think this is the historically strongest thrust in terms of its trending properties, but this may be the best organized ones to get the most bang for the buck and I think the thrusts from the 2009 lows were stronger than this one.

I think you're addressing something that should already be understood (but perhaps I'm being presumptuous): Beyond a certain extreme point in breadth, a further extreme in breadth may not necessarily lead to better performance going forward...indeed, there is no consistent proof that it does, and there are too many other variables to account for, in any case. This is why I did not indicate anywhere or suggest that this breadth thrust will perform better or worse than the others.

Just have a look at the table and you will see that some of the best performers are at the bottom of the table. All of them were above a threshold of 75, however, which represents an extreme reading.