The brokers numbers are based off the implied volatilities, so i dont think one could be said to be wrong and the other right, only that a hypothesis cannot be rejected or accepted within a certain confidence interval
Risk-Neutral Probabilitiy: Intuitive explanation?
- Thread starter stochastix
- Start date
let me ask you a very simple question genius and im sure you will google it.
what about goodness of fit what exactly do you think about that with your curve fit model
what about goodness of fit what exactly do you think about that with your curve fit model
I think its an interesting question.. when you are fitting probability distributions, its not called a curve fit since its a stochastic model, its called a calibration and estimation. Curve fitting would be when you do something naive like fit a polynomial/Bezeire curve to a price series and try to extrapolate into the future with no economic model or ideas behind it.Trying to make point estimates or something
So yes, model could be miscalibrated or misspecified, but its not, its a pretty good fit, there are other better fitting models that I havent implemented yet. When you study statistics you do maximum likelihood inference, and it lets you compare hypothesis .
So yeah, you need to calibrate your model on the appropriately chosen length of data, or find some way to fuse data into estimates as the data happens (Kalman filtering and the extensions thereof)
Do you even stastics? Have you ever taken a wad of numbers and found out what distribution describes it best out of a set of alternatives?
You know you can take an emperical histogram, and then show the theoretical probabiltiies, and the closer they match the better the model is right?
blah blah blah where was your model on this sell off? just because you fit a model with too many paramaters and nailed an option trade to the T doesnt mean it is going to work tomorrow or next week that is my point.
obviously you must keep searching and tweaking and that was my point from the very beginning. what models and works perfectly right now or yesterday is still a best guess for tomorrow
obviously you must keep searching and tweaking and that was my point from the very beginning. what models and works perfectly right now or yesterday is still a best guess for tomorrow
trader, yes, its always a moving target. the model worked perfectly fine on this selloff and I had accounted for the possibility beforehand due to what the model indicated
after stepping away why are we argueing shit we should all be trying to work together but that is the state of where are now huh all of us would rather argue and fight and it smy way not your way and its even apparent on this site over and over.
agreed always a moving target totally
Thats a deep and good question, I do not know. I think this has a lot of explanatory power though
https://www.independent.co.uk/news/...a-thin-line-between-love-and-hate-976901.html
the pathways for love and hate in the brain are largely the same, its just how other pathways are expressed simultaneously that depends on how these are expressed.
"They found that the hate circuit includes parts of the brain called the putamen and the insula, found in the sub-cortex of the organ. The putamen is already known to be involved in the perception of contempt and disgust and may also be part of the motor system involved in movement and action."
Even when I worked on wall-street, I had a boss that was so wealthy I wasnt allowed to speak to him because the subordinates feared that "ANYTHING!!!!" could set him off. The dude was mega-wealthy, but mad, and spent almost every day clamoring to get into this office in lower-manhattan at the ass crack of dawn for no good reason, and everyone complained about "not having any time" . They paid more on their monthly office lease than my yearly salary, but would fight you for weeks to get reimbursed for basic necessities and office supplies.
If you want to use the CIR model yourself, there is matlab code to estimate the parameters , https://www.researchgate.net/public...ersoll-Ross_Process_The_Matlab_Implementation . apparently there is a lot of variance in the estimation of the risk premium, lambda, so it is put in by hand usually