Quartz

326 reputation

bio	website
	location
	age
visits	member for	7 months
	seen	May 17 at 12:47
stats	profile views	19

Monte Carlo, risk, QMC, statistical efficiency, high dimensional approximation...

	326 reputation	bio	website		visits	member for	7 months
6 badges		location			seen	May 17 at 12:47

summary answers questions tags badges favorites bounties reputation activity

51 Actions

suggestions reviews revisions comments badges posts accepts all

Feb 15	comment	Why is C++ still a very popular language in quantitative finance? Not in the current project since they're still stuck with legacy code, but I've used it with full satisfaction in the past (and plan to do it again whenever possible). I also know some people deep in the HPC linear algebra R&D world, and their way more competent appreciation is significant. Unfortunately quants cannot always carry out a throughout comparison before adopting a lib (and many dont even have the numerical background to do so), so historical reasons dominate technical considerations, but my impression is nevertheless of slow but steady adoption growth.
Feb 13	revised	Why is C++ still a very popular language in quantitative finance? added 93 characters in body
Feb 13	answered	Why is C++ still a very popular language in quantitative finance?
Feb 13	comment	Why is C++ still a very popular language in quantitative finance? Ah, I forgot, some SSE instructions even specifically target memory performance optimization!
Feb 13	comment	Why is C++ still a very popular language in quantitative finance? 2: see my comment above on GC: you can do GC in C++ if you want, while you can't easily avoid it in Java. C++ lets you free to chose according to the situation, while Java forces a suboptimal allweather compromise (also in other cases, such as unsigned). \\ 3: red herring: again memory usage is surely critical, but that's true in all languages, and the ability to additionally optimize calculations is a clear plus. Impeding FP performance optimization does not improve memory usage in any way, so why should that be good?
Feb 13	comment	Why is C++ still a very popular language in quantitative finance? You can have garbage collection in C++ too, so it can't be cited as an advantage of Java. Java often forces an unflexible paradigm on the user on the ground that "they know better", while C++ lets you free to adapt wisely. Unsurprisingly Java was repeatedly forced to backstep and readopt features of C++. Security of Java also turned out to be vastly exhaggerated.
Feb 8	comment	Discrete returns versus log returns of assets There's a further reference from Meucci specifically on this latter procedure: papers.ssrn.com/sol3/papers.cfm?abstract_id=1586656
Feb 6	awarded	Commentator
Feb 6	comment	How to simulate one-minute bars data from one-day bars? Sure. And the devil is in the details: you also need the conditional joint PDF of where Open & Close fall :-)
Feb 6	comment	How to simulate one-minute bars data from one-day bars? @Joshua Ulrich: What if your simulated path is concave or convex while target max and min lie on different sides? Sure, one might reject until this avoided, but it was just to show the problem... again, what if one simulated extreme is on the correct side but very near open or close, while open/close are far from the line? The resulting process would behave in a very strange way and introduce artifacts hard to control. Unfortunately one really has to condition on max and min too, which is not that trivial. Maybe a smoother transformation and rejection threshold might be a crude hack though.
Feb 5	comment	How to simulate one-minute bars data from one-day bars? @Richard: of course BB preserves open and close, but if later a scaling is added that will be lost.
Jan 31	comment	Consistency of economic scenarios in nested stochastics simulation Sorry, could you please state the consistency part more precisely, to better agree on terms and the problem? Anyway I do not see well what's the issue: at the different RW scenarios in t=1 you calibrate and price with a corresponding RNt1_i measure, which sure will all be somehow related to RNt0 by construction, but why care about that?
Jan 22	answered	Overview of software companies in the industry
Jan 22	comment	Overview of software companies in the industry It's already hard enough finding a comprehensive sw list, let alone a comparison; for a partial list check bobsguide's directory. For comparisons Lupus offers something, although not for free.
Jan 21	asked	Basel CVA VaR with R/WWR
Jan 21	comment	How to simulate one-minute bars data from one-day bars? So you dont need open and close to be preserved?
Dec 13	comment	Is it possible to model general wrong way risk via concentration risk? Could you please explain a bit further what you mean? Of course there can be WWR even with a single market factor, while on the other hand WWR will interact with any market risk. So WWR on concentration risk is necessary but not sufficient, but this is probably not what you're asking for, right?
Dec 10	revised	Model-implied yield spread on corporate bonds Cleared adherence to question.
Dec 10	answered	Model-implied yield spread on corporate bonds
Nov 26	comment	Reference request: Survey article on GPU in Finance I havent found "official" statements of that bank's decision, so it's just a rumour that would be unsafe to spread further. But it can be asked on the LinkedIn group (which is "GPUs in Finance"), many guys from such bank are reading so I leave them the choice of wether to share their experience on electronic ink. And I know of no good survey, only specific applications (which would be OT here).