Having internalized some basic aspects of wave counts, such as alternation of corrective waves within a motive wave, coming back to the counts produced by Advanced GET is a strange experience, as the software-generated counts seem quite wrong.

Have others, as I now have, given up using software to mark the key wave points? Of course one would still use a software grid to mark Fibonacci retracements.

Anatoly Veltman writes: 

Actually, Advanced Get by Tom Joseph was very good when first introduced in late 80's-early 90's. Trick was that one should have also attended Tom's weekend workshop (mostly held near an airport in Ohio), to be tipped on the whole essence: type 1 and type 2 trades, wave 4 index and oscilator. Without figuring out when Wave 4's odds diminish to unacceptable — there is no reliable Elliott Wave trading. And Fib retracements are great — but ONLY if EW type 1 or type 2 trade has first been isolated. I taught Tom's methods for about 15 years. Not sure if any of my students succeeded in black-boxing the entire methodology.

Tim Melvin writes:

Did someone really say fibonacci on the spec list? This could get interesting if it is anything like the old days…

Anatoly Veltman writes: 

Well, that's the whole point. Loving to say Fib doesn't test well– when the wrong application was tested to begin with.

Phil McDonnell writes: 

To be sure one must test something according to the right way of doing things. However that is exactly the problem with wave counts and the like. The rules are so arcane and convoluted even so called experts disagree on them.

If you get 5 different Elliot exerts in a room you will get 5 different wave counts at the same time. It is a bit like the game of Fizzbin. The rules keep changing and are unnecessarily complex. 

Leo Jia writes: 

I think one probably should take this argument as a not-bad news for Elliot theory or any theory that gives non-consenting results. It means that it likely has some statistical truth in it that is worth one's effort in seeking. Don't we agree that a market theory delivering definitive results does not exist or, if exists, ought to be thrown out?

Steve Ellison writes: 

Trying to stay in line with our raison d'etre, I have been coding a method for retrospectively identifying highs and lows of multiple levels of significance.

My approach is to go bottom up, starting with an idea I got from one of the Senator's books. A local high is a bar whose close is higher than the closes of both the previous bar and the following bar. A local low is a bar whose close is lower than the closes of both the previous bar and the following bar (a sequence of 2 or more bars with equal closes count as one bar for this purpose).

After identifying the local highs and lows, I move up a level. A 2nd level high is one that is higher than both the preceding local high and the following local high. A 2nd level high cannot be recognized until one bar after the lower local high that follows the 2nd level high. I record the time at which the 2nd level high could have been recognized.

I follow similar rules to identify 3rd level, 4th level, etc., highs and lows and the times at which they could have been recognized in retrospect.

I haven't finished yet, but this method should give me a platform for testing hypotheses about "primary trends", etc.

Anatoly Veltman writes:

Tom Joseph's contribution to E.W. trading, in my view, was much greater than Prechter's or RN.Elliott's. Tom basically said with his excellent refined Type 1 trade: don't ever place any bid, unless:

1) you've already observed a valid impulse (with extended third wave)
2) a correction is currently in progress, approaching 38% of preceding rally
3) you're filtering this correction with oscilator return to 0, and fourth-wave index still sufficient for fifth wave
4) fifth wave projection extends to at least 2:1 profit/loss ratio, incl. all possible slippage.

I say: if all these conditions are not met (and this may not occur every day) - never place a bid at 38% retracement. If all these conditions are not met, you'll have to bid only at near-100% retracement. What does this principle have to do with popular E.W. or popular Fibonacci methods. Nothing!!

Laurence Glazier writes: 

Sure, things are complicated and one would not wish to poke a stick into a hornets nest, but … some things are complicated.

It took hundreds of years to elicit the laws of harmony from the canon of classical music (many to this day deny their existence). Put five composers in a room and have them harmonise a tune (the non-believers might refuse to!), and they will do it five different ways, but they will all have added to the map of knowledge.

Even knowing those laws, one could not reasonably predict how a piece of music would continue if Pause were pressed (unless it were minimalist) - but one might anticipate it would return to the tonic key, and that the free fantasia would not be over-long, and so on.

Those laws are difficult, unprovable, and without material substance but are the result of empirical observation.

Gibbons Burke writes: 

CTA E.W. Dreiss used, in the 1990s, a very similar way to count waves in the market using what he called the Fractal Wave Algorithm (FWA), and he traded futures breakouts from FWA-n magnitude highs and lows. Did quite well, but like all trend followers, it is a bumpy ride.

He also came up with the Choppiness Index, which sums the true ranges in the last n periods, and takes that as a ratio of the n-day range.

Jason Ruspini writes: 

This is the natural approach that I took as well. Ignoring the "correct" 1-5 definitions, I just looked for a run of higher such double-X highs and higher double-X lows identifiable with the necessary lag, with attention to what happens when you eventually get a lower major high/low, breaking the "wave" run count, which can keep going after 5. What I found wasn't very interesting, in-line with my previous comment. I'm still unclear if anyone is actually trading a tested (complicated) system or just applying versions of rules with discretion. If it is a tested system, why is it better than a simple long-term momentum system?

George Parkanyi writes:

I like to keep it simple. Many years ago, I read something written by Larry that said, when the commercials are generally substantially more net long or short than specs - that tends to stop trends and turn markets the other way. He admitted it was a rough rule of thumb - that it may take a while to turn the tanker - but I pay attention and time after time I've got to say it works. So right now two markets that fit that profile are coffee and to a little lesser extent sugar. (Oh yeah, VIX as well) I've been long both for a couple of weeks with modest starting positions, and just had a nibble at VIX. I don't know when the trends will turn and I may have to take a stop or two, but I like the chances for a good position-trade in these two markets - and VIX as a bet on a short-term post-Fed hang-over. I checked back to when coffee started this particular big decline - and it was within two weeks of when commercials were selling the crap out of it and their net-short positions had peaked. Gold and a number of other commodities did the same thing at the beginning of this rally that began in May - except that the commercials were the only buyers at the time. It may be a dumb-as-dirt perspective on my part, and will likely set off Anatoly - but its one thing that has stuck with me from reading a number of Larry's books.





Speak your mind


Resources & Links