A Kinetic Energy Idea

maestro-wondering what software you are running on your 24 monitors -if indeed you are that person. thanks

 

--------------------------------------------------------------------------------
Quote from alex.samant:

i would probably do fine reading the tape but i'd still have to draw price action on a piece of paper because the brain can't remember so many numbers.... the only simple reason i guess...
--------------------------------------------------------------------------------


M's reply:

Exactly! So if you had the means of helping the "tape reading or chart reading" without substitution of the “raw” info on it would that be better as well? You see, the point I am trying to make is that using helpers (models, indicators, goggles, rulers etc.) is to enhance the ability to process the otherwise "hidden" info that escapes our natural abilities to comprehend.

Make the right 1/3 of your chart the future. Annotate there. you can fill in the next three trades in terms of the locus (time, price).

The modern substitute for reading the tape is the mathematical combo (script) of the T&S, The DOM, PRV and the Premium. These are also usually displayed by expert traders. As you well know the future is showing in these data. It is the equivalent of the blank right side of the price chart that you are annotating except the market is annotaing the future in both price and intention at that price.

We all know that price "turns" at the wall on the DOM and it gives us the OPPOSITE of what the guy with the ball and the incline thinks he is seeing. The market is counter intuitive. Pricce moves because the minority goes to zero and all those remaining in the majority are simply passed by. LOL

You can look at the drift of the premium continually and superimposed on the drift is a leading indicator of price in the form of the Connors- Hayward volatility compression and then the END of the compression with what is known as volatility expansion of the _________!!! How long before the wall is hit and the turn occurs does this lead price???

How long will it be before you have a valid relationship between market pace and volatility? Right now, you haven't an inkling of what trumps what and when. LOL.

Have a nice holiday and reconsider your name calling. Exactly.! .... LOL....

 

I wouldn't say i miss it Bob. I am working on some feed forward NN's and i am finding that normalizing the data is the hardest part. Because although there is only the bid/ask and there sizes (open book etc...) there is so much more to it than that i.e. who is going to move the market and for what reason. I just recently started calling it Psychologicompting thanks to the article recommended by nitro.

Yes, the physical pit is moving towards extinction (pretty sad to see the spoo pit at the CBOE compared to what it once was), but the market makers and specialists are still alive and kicking. New structure, same purpose: Liquidity.

 

I use NNs exclusively for all aspects of trading. Yes, data per-processing (normalization) and post processing are really hard parts. Most traders who lack a formal math and IT background fail at this miserably and conclude that soft computing doesn't work.

The other part that requires significant intellectual insight is the architecture of the Independent Variables. If you just toss 20 traditional Technical Indicators at an NN you’ll find it doesn’t work much better than trading the TIs manually since the traditional TIs were designed for manual use and ease of calculation and understanding, not because they capture underlying market dynamics that have predictive potential.

However, I find that price alone (transformed into Neuro-TIs gives all that is needed. Who is doing the trading doesn't matter or perhaps it's reflected in price anyway.

Jerry030

 

This is for MAESTRO’s original post.

1.Market is in constant moving due to non-stop trades, the source of energy. It is a great idea
using kinetic energy to model the market movement.

2.Any indicator should be normalized which include your newly created kinetic energy. Otherwise
the computer cannot use it to make judgments. This is model independent, regardless of friction
move or frictionless move. It look like you are confused and mixed these two concepts.
Nevertheless, the normalization idea is correct.

3.Why use ATR to normalize k_energy? These are two different physical quantities. ATR measures
the RANGE (Max & Min) but does not tell WHEN to reach them. On the other hand the k-energy
measures the rate change of speed for fixed time frames (WHEN) but does not tell how far (WHERE)
the market had been away (max/min) during the time frame. One way to normalize it is to use AKE
itself (I call it for Average Kinetic Energy) to replace ATR, in past 14 bars if one prefer in your
formula. To completely normalize it you can use formula of

(K-E of last bar – min K-E in last n-bars) / (max K-E in last n-bars – min K-E in last n-bars).

It will give you a 0% to 100% K-E indicator.

4.There are much more works to be done if you want build K-E model with friction. It is non-linear
as the friction increases for higher speed (market moves faster). In addition, how to distinguish
friction volumes (against moving direction) from active volumes (along the moving direction)?

5.K-Energy only models partial energy changes in system. How about the potential energy?
You may be able to figure out what P-E model to use (definitely not the P-E from gravitational field).
These will give a more complete picture. Please note that even so, the energy conservation theory
does not hold true here because markets are open systems rather than close one.

 

agreed sometimes i feel i have Asperger syndrome when it comes to reading tone in posts. maby i can develop a FFANN that derives a posts intention through lexical analysis.