Mostly messing with Brownian motion this evening - I've been trying particle swarm stuff too but I keep fucking it up so stay tuned.
Never once touched Brownian motion, but I know it's used in biophysics a lot. What would you use it for?
Short term asset pricing and prediction.
BM is the foundation Black-scholes options pricing along with many other pricing models that utilize the efficient market hypothesis.
I don't really believe in EMH personally and foresee a revolution in modeling when the old guard dies off, but its usefulness in short term intervals is undeniably useful.
I gather the particles are a 2D simulation of springloaded balls in zero gravity so when one is pushed it'll collide with the others and the others to eachother and over time it becomes more active and intense.
Other than that I have little to no idea how you'll apply information from that.
I gather the particles are a 2D simulation of springloaded balls in zero gravity so when one is pushed it'll collide with the others and the others to eachother and over time it becomes more active and intense.
Other than that I have little to no idea how you'll apply information from that.
No a closed system with a constant energy undergoing diffusion to maintain equilibrium - modeled using Wiener Processes and Boltzman equation.
The distribution tracks that diffusion over time, notice that despite seeing singular spikes outside of the curve it constantly remains below the curve in the aggregate.
These same principals can be used to price assets and predict returns because if one assumes the validity of the efficient market hypothesis then price dynamics are said to be random. Weiner Processes are a means to model those random dynamics and as such can give one predictive power by decreasing the amount of uncertainty you have about those price movements.
Essentially kinetic energy of the system applies to volatility which can be modeled using diffusion. I can use diffusion to model price movements because despite being a random process at the micro level -price fluctuations- it as a whole informs an easily predictable macro trend (the equilibrium of the closed system).
All I want to do with this is predict mean reversions which equates to identifying spikes, the probability of those spikes, and in their return to normality to maintain equilibrium.
I was thinking kinetic energy but in 0 gravity since the speed seems to remain the same until they make contact.
Still don't understand it, but if you manage to make an excellent indicator, don't forget me, Jew.
Of course I won't forget you, you've helped me so much.
As soon as i get this down I'm going to apply it to crypto as well, I think my account will be where I want it by the end of the month that I'll put 10k-20k into binance and start building a crypto position - don't want to miss the big move when it happens.
This morning I was messing around with particle swarms as a means of optimization, this is a really simple example of a swarm tendoing across a negative gradient searching for a minimum.
Going to start messing with maps with local minimums now.
For Sir Tony, this method of optimization can be used in portfolio creation - gives you better ideas of how to minimize risk and maximize returns and yield.