Regarding the no boundary proposal: Ideas are useful instruments, and the worth of an idea is based on how effective it is in explaining and predicting natural phenomena .
Math has to be applicable as in physics, if you go beyond that it is just abstractions that have no basis in reality.
There's obviously quite a lot of mathematics that indeed has no basis in reality is just sort of in reality in that the people who invented it were part of reality but it doesn't describe I've read the gamut of it it takes a lot to know what's useful got to go through a lot
Well I guess volatility models are part physics then cuz they certainly do a great job at describing the variation of random phenomena
How are volatility models correlated to observable reality ? You are trying to solve problems from the starting point of equations and mathematical structures.
That's the only way to solve anything from first principles ab inito. If the model can perfectly calibrate and give prices that fall within the bid ask spread for the thing then it effectively describes the stochastic process of the s&p 500 and Vix. You can do no better than that because it's effectively the same as predicting the distribution of possibilities of the position of a particle which cannot be predicted with certainty. This is apparent when you study the canonical commutation relations in either the Weyl, Heisenberg or Schrodinger form. The observables here are the prices and the model I'm implementing does indeed perfectly calibrate and it's derived from actually a microstructure Hawkrs process model which is a self-exciting model that actually does a pretty good job of describing the pertinent features of the limit order book Dynamics and they actually show the dominated convergence very rigorous arguments that this microstructure model converges macroscopically to a rough fractional volatility model that is a volatility model with a fractional exponent denoting long memory this exponent is around 0.1 for most financial markets. The Hurst exponent - a perfectly calibrating volatility model is indeed "as good as it gets" in terms of describing the underlying stochastic process. Here's a more accurate way to think about it: 1. A perfectly calibrating model, by definition, matches all observable market prices across strikes and maturities. 2. These market prices embody all available information about the underlying asset's behavior, including expectations about future volatility, skew, and other stochastic properties. 3. these prices represent the best collective estimate of the asset's true dynamics. 4. The lack of deterministic prediction is inherent to the nature of stochastic processes. The model captures the probabilistic nature of the asset's behavior, which is precisely what we want. 5. The model's ability to perfectly replicate market prices means it's capturing the market's aggregate view of the asset's stochastic properties, which is the closest approximation to "true" dynamics that we can hope to achieve. 6. Any deviation from this perfectly calibrating model would imply that we believe we have better information than the entire market, which is a very strong claim. - a perfectly calibrating model is indeed capturing the dynamics of the stochastic process faithfully, insofar as we can define and observe them. The non-deterministic nature of the predictions is a feature, not a bug, as it accurately reflects the inherent uncertainty in financial markets.
Still if your starting point is of equations and mathematical structures it reveals nothing known either true or false about nature's unobservable objects. Hence math only describes and never explain. Wouldn't that make you yet another shade of TA trader ?
A benefit of having a model is you can see how the parameters of the model changed over time rather than just observing the stochastic process. These parameters are in the quadratic Heston directly interpetable in terms of the minimal instantaneous variance, The asymmetry between reactions to positive versus negative price movements, and the feedback that models the Weak and strong zumbach effects and of course the Hurst exponent and the volatility of volatility. If you have a view that some of these parameters are going to change in the future you can actually put that into your model once you've calibrated it and see how that prediction deviates from the market predictions for instance
No it wouldn't because TA traders are like trying to talk to schizophrenics it makes about as much sense. Nothing is done in a rigorous stochastic analysis paper that doesn't make sense and what doesn't make sense the try to denote it as such but I've never seen any rigorous justification for any crap any TA BS artist has spouted