Quote from maxpi:
I don't have the grail, I have just progressed to where I was able to chuck a ton of random junk on the random junk pile and get on with things and even maybe understand what Jack is talking about a bit as well. I am pushing towards something very simple that can be automated however and am quite happy with the progress on that. I will say that the advent of the personal computer has brought a ton of confusion to traders because it is easy to make indicators that use nothing but price data to predict price, and mathematically, that is not possible. Wyckoff wrote about price interpreted by volume a long time ago and Toby Crabel wrote about volatility predicting volatility a couple of decades back, reading those two authors is a good idea.
I'm exploring an odd thing currently that is raising questions for me. If I offset the begin time of all the bars by offsetting the begin time of the day it has an effect on the overall outcome. I also find that I'm using essentially two systems overlaid and when the offset takes one of them down it improves the other... curioser and curioser..
Super post....
Quite a while ago I used a drafting board to plot data during the day. Then, I was carving turns at about the 10 to 15 rversals a day. The offsetting of sets of data is one of the best ways to combine the Wyckoff notions and the Crabel notions you mentioned.
The preprinted grid I used was for 30 minute bars overlapped by 15 minutes. What you see is each 30 minute bar occupying half of the time period in width. The 26 slices is double the # of 30 min bars in a day during RTH's.
I projected the cross hatching to be able to stay at least 15 minutes ahead of the markets for turns. The last two bars were used to project. You can see the color coding and the lines used to project. Yellow started the day and the orange kicked in 15 minutes later. This was when YM was in the 9200 range as you see.
When I started with commodities there was no YM it was only DJXX then. The YM added a lot of volatility and volume thank goodness.
I saved boxes of these kinds of charts in with the other daily prelight and debriefing.
What I learned then was to keep slipping down to a finer measure of the markets. Nowadays it is easy to go out ahead at least two trades but more than P and V are used. The accuracy is to the finest measure of the market (a tic).
One of the things to really consider is that not all traders gt what they want. A lot of order that show are left on the table when price changes.
For example, if you leave price out of the picture, you can use volume equations to do all trading turns. But it is not necessary to leave out price as a lagging indicator of volume.
When you look at the markets in terms of the market operation and the trader operation in terms of trader plays separately, you get to always see a phantom screen there that displays what is going to happen fairly soon.
The phantom screen is filled in with the price and time of at least two trades out. Sometimes in a trading room I put my finger on this place on a chart. People then get to se it happen.
I found that it is possible to get to simplicity. My path didn't go there directly. First, I had to go to comprehensive. I do what you are doing:throwing things away. mcichocki is doing the reverse: he weeds before he gets to see what is possible to see.
He feels Wyckoff has it bawards as he states.
Now we have the PC and coding available; it works faster than we do by 5 to 20 times. That makes slicing and overlapping so much easier. And we can use more data too.
It gets very simple after all.
