Monthly Archives: August 2013

Thoughts on “wave analysis”

Here we go. After calculating extremes, the next point is to calculate and track how price moves to extremes. These are the steps to create waves. Incredibly important. If the waves are constructed incorrectly, the pattern will not show. The nice thing is I can trial and error my way to getting “correct” waves, given that I am viewing data in the correct light.

There are 2 parts: Distinguishing waves, and then finding ways to get excel on the same page. Here are the 3 types of “waves” I think I want to try and start with.

Quick ThSpeaking only for bull movements, price through the day, respective to open, can:

1. Move straight up, hit an extreme, and then retrace a bit.

2. Retrace first, then make the extreme.

3. Make the extreme, then retrace, create a new low, but not lower than the length of the bull extreme.

Put into other phrasing: The extreme high of the day occurs at some point. The extreme low can occur at 3 points, at the open, before the high, or after the high.

It is possible to come up with many more waves, which may be needed. I may only need 4 types of waves (total) instead of 6. Maybe more. Ideas from Rel were flats, retrace into new highs, and others. Lets see if I can start simple first.


Basic movements into Extremes

I wanted to take a quick glance at movement into the extreme. I looked at basically 2 things. In each day, relative to open, is the high bigger or is the low bigger? If so, which comes first? If we assume the bigger movement is the “correct” trade decision, the question is more like asking “will there be a dip or retracement before the big move? Or is it more likely to retrace after the extreme is reached?”

First I recalculated the min/max points, as well made two new columns for how “big” the ups and downs were. Which was more extreme, the high or the low? Then using vlookup and attaching numbers to the hours, I calculated which came first in the day, the high or the low? Finally, I checked to see if they were the same: “is the bigger also the same as the first?”


And the results:


A bit surprising. And convincing. If the up move is bigger, the low will occur first, and if the down move is bigger, the up move will occur first with 80% accuracy. This seems like it has potential.

Question: would the results be different if I calculated the bigger move as high-low of the current day rather than high-open of the day? (for up moves and vice versa or down moves)

Dear Journal #2

Dear Journal,

It’s been a while. It’s not that I haven’t been trying, it’s that I’ve been stuck, among dealing with quitting, moving, and other life things. Most importantly, stuck. How to make progress?

There are still the old important questions, with a few new ones, that need to be answered:

1. How do I deal with the issue of multiple time frames? Trends in one time frame are retracements in larger time frames. In some similarity, trends and retracements in one time frame can be considered noise in a larger time frame. High tier scalpers and large intraday traders care a great deal about getting the absolute best price when the magnitude is millions of dollars worth. The pip and the pipette matters. Big position traders, who really run the markets, are slightly different. A pip or 2 off won’t kill a position that is set to make well over 200 pips anyway. Tying this together: What can be considered a big S/R watch zone for small tf players isn’t the same for large tf players. When orders come in from big money, price shoots. Back to what SB said, it’s important to know which tf is in control, and when. To reiterate what Rel said, Small tfs are more time sensitive, big tfs are more price sensitive.

I’ve been told pure intraday data can be good enough, or fairly close. Weekly and monthly trends, due to their nature, have greater variance. For intraday trading, trusting the intraday trend more than the weekly trend is logical. I think it would be nice to find a way to tie in waves on one tf into the waves of the larger. Easy to think, probably a monster load to do.

2. How do I deal with the issue of “pip difference”? Daunting. Extremely. A pip in 2000 is not the same as a pip in 2013. How can I map price movement consistently across years? Considering Rel said even he couldn’t find a solution and that it was given to him, I fear for my ability to solve this. “I never would have thought of it”.

3. How to make progress on the wave indicators?

This is the most practical or hands on of the 3 questions at the moment. How do I actually map price movement? The main idea is to keep things simple, yet capture the necessary action. Drill drill drill. Start with mapping extremes, then map how price moves to extremes. Accumulate statistics for findings along the way. The idea is to monitor a sort of “life cycle” MMLC (major move life cycle) of price movement. I’m told eventually I want a sort of “scatter plot” that will do something for me… What benefits does the scatter plot offer over my traditional %s? Can I do without it?

Time for Daily Extremes

As I think it should be better called: Day extreme Statistic

I think there is a bit of a problem with calling days based on Oanda days, for 00:00 EST days. With an FX mindset, it’s problematic to call the end of a day arbitrary. It’s worst that 00:00 happens to be in the middle of the Asian session. If each session really did have it’s own wave, then that is definitely an issue. I’ve always preferred to consider day starts and ends at the beginning of the London session because that’s when the volatility ramps up, but that’s also in the middle of the Asian session which is low volatility, but not totally dead. What’s totally dead? The gap time between the end of NY and the beginning of Asian. So I called 20:00 EST (Asian open) the start of the day. Then I re-ran the statistics. What time do the Extremes occur?


Statistics don’t lie? The data is still presented in EST. It appears Asian open is actually the extreme of the day more so than any other time. Other numbers make some sense. Late Asian-Pre London Open sometimes presents an extreme, The hour before and after NY Open tends to be a good place for extremes, and extremes tend to drop off after London Close. Nice! The major problem however is that the statistics are much more flat. Nothing stands over 10%. When statistics are run London to London, things are much different. Peaks appear much more easily, and are much more pronounced: