Finance data analysis_nuclear_physics_comparaison
0 likes508 views
This document compares nuclear physics and quantitative finance. It outlines several similarities in their strategies, modeling of fluctuation/randomness, use of mathematical modeling, and potential for catastrophic outcomes. Both fields gather data from experiments/simulations, perform quantitative analysis using programming/platforms, and aim to predict outcomes. They apply similar probabilistic, random walk, and option pricing models. While also noting differences like physics' conservation laws and reproducibility, the document questions if finance has violated physical laws by prioritizing unsustainable growth. Overall, it finds quantitative parallels between the two domains.
Finance data analysis_nuclear_physics_comparaison
- 1. Comparison Between Nuclear Physics and Quantitative Finance Ahmed Rebai, ahmed.rebai2@gmail.com PhD in Nuclear Physics
- 2. Warning and risk disclaimer Nowadays Trading carries a high level of risk, and may not be suitable for all investors. Since I am not an economist, an arbitrager, a market analyst, a market bull, a broker, or even a day trader (for the moment), I disown any responsibility for any errors or misunderstandings caused by this presentation 1
- 3. PLAN Definitions Similarities: Strategy (Data=>Analysis=>Results) Similarities: Fluctuation and Randomness Similarities: Mathematical Modelling Similarities: Big Catastrophes Some Differences Did finance violate the laws of physics? Conclusion 2
- 4. SOME DEFINITIONS Quantitative Finance: use quantitative techniques like mathematics, statistics, computer science to try to model the value of financial securities and structure them to hedge them. Deal with some questions about:How interest rate be in the futur ? How volatility be in the futur? Nuclear Physics: study subatomic structure of atoms and nucleons using large experiments, mathematical models and statistics techniques 3
- 5. DATA ANALYSIS RESULTS Both are using the same strategies SIMILARITIES: SAME STRATEGIES 4
- 6. DATA 5
- 7. DATA (PHYSICS) Simulation: Monte carlo Generator Data Nuclear physics FLUKA, MCNPX, ... Accelerator experiment Particle physics Pythia, Geant4, ... Diffusion experiments Astroparticle physics CORSIKA, AIRES, REAS3, ... Cosmic rays experiments Data is often generated by experiments or by monte carlo generators. (A Monte Carlo technique: is any technique making use of random numbers to solve a problem.) 6
- 8. DATA (FINANCE) Special sites: Blommberg, CNBC, Reuters, ex.com, DailyFx.com, ... Economic official report : Non Farm payroll (NFP), UK CPI, German GDP, ... Economic indicator: S&P500, NASDAQ, DOWJONES, FTSE … Using Monte Carlo to generate data. 7
- 9. ANALYSIS 8
- 10. ANALYSIS (PHYSICS) Data Analysis needs computer resources: programming languages + infrastructures Low level languages: C/C++ (open source) High level languages and Platforms: Python, IDL, Matlab, ROOT-CERN Linux Environment... GPU/parallel programming: CUDA (NVIDIA) , MPI, OpenMP, ... 9
- 11. ANALYSIS (FINANCE) Fundamental analysis take much time, Technical analysis no so sophisticated, Quantitative analysis takes into account market evolution and its random nature. Then data analysis => pricing proces (pricing, Forecasting, prediction...) Needs dedicated platforms: Metatrader, Ninjatrader, Tradestation... High Frequency Trading => using algorithms C/C++. 10
- 12. RESULTS 11
- 13. RESULTS (PHYSICS) Study the fundamental laws of nature. Search predicted particles by the standard model (eg the Higgs boson) Or new particles (predicted by supersymmetry theories beyond standard model) Exploring the universe (dark matter, dark energy) 12
- 14. RESULTS (FINANCE) Predicting the market movement Make a good return Profit Money $, €, ... 13
- 15. SIMILARITIES (FLUCTUATION & RANDOMNESS) In nuclear physics: arising from ” the Heisenberg's uncertainty principle” (Quantum nature of laws in atomic and subatomic scales): => Can't predict simultaneously the position x and momentum p of particles. In quantitative finance: arising from ”the efficient-market hypothesis”: => Can't predict the future market volatility and price... 14
- 16. SIMILARITIES (FLUCTUATION & RANDOMNESS) In finance: the only reality is that price may go up or down (increase or decrease): In quantum physics: quantum state involves a superposition of quantum states of 2 different particles ((Schrodinger's cat paradox): alive dead dead 15
- 18. Mathematical Model $ $ $ $ $ $ $ $ $ $ $ 17
- 19. A fluctuating stock price => need for probabilistic models => Random walk model “Today’s price (PT ) = yesterday’s price (PT-1 ) + a change that is independent of all previous information (CT ).” PT = PT-1 + CT (Since P1 = C1 P2 = P1 + C2 , P3 = P2 + C3 => PT = PT-1 + CT) The most important model for equities, currencies, commodities, bonds and indices. SIMILARITIES: Mathematical Modelling 18
- 20. SIMILARITIES: Mathematical Modelling The Cox-Ross-Rubinstein option princing model Uses a discrete-time In a perfectly efficient market:No possibility of arbitrage, At each time, the price can increase or decrease and never both simultaneously. 19
- 21. The Black and Scholes model ”The history of quants on Wall street is the history of the ways in which practitioners and academics have refined and extended the Black-Scholes model” Emanuel Derman's book (My Life as a Quant: Reflections on Physics and Finance) SIMILARITIES: Mathematical Modelling 20
- 22. The Black and Scholes equation can be transformed to a heat equation model: Many methods of resolution: Green's function formalism, Numerical resolution (e.g finite differences)... SIMILARITIES: Mathematical Modelling 21
- 23. SIMILARITIES: Mathematical Modelling In physics to study the random motion of a free particle in space (phase space)=> Wiener process => Diffusion => heat equation model More general model : Fokker-Planck equation (Diffusion and convection) In same case FK equation can be transformed to a heat equation (cosmic rays acceleration) I'll stop here because things start to be really complicated !!! 22
- 24. But when things go wrong: In finance: Global Financial Crisis, Flash Crash, Loss of money, Recession... In nuclear physics: Nuclear Holocaust: Hiroshima, Nagasaki Nuclear Incidents: Chernobyl, Fukushima Similarities: Big Catastrophes 23
- 27. Similarities: Big Catastrophes 2010 Flash Crash 2008 US Crisis 2008 Europe Crisis 26
- 29. SOME DIFFERENCES In Finance you are dealing with people ”the crowd psychology” The world of finance and the world of people is changing all over the time... History doesn't repeated itself.... Where in physics history repeats itself all the time you can do the same experiment over and over again (Reproductibility) In physics you deal with particles, atoms... 28
- 30. In physics conservation laws: electric charge conservation, momentum conservation, energy conservation In finance no conservation laws: There is no conservation law in stock market but it depends on economic model... SOME DIFFERENCES 29
- 31. DID FINANCE VIOLATE THE LAWS OF PHYSICS ? Bankruptcy of Lehman Brothers The 2008 Financial Crisis 30
- 32. DID FINANCE VIOLATE THE LAWS OF PHYSICS ? The myth of economic exponential growth model meets the finite physics ressources : Finance have long been inconsistent with physics laws especially since the financial crisis unbalances the instability of the global financial system.. But why ??? 31