I can't help asking this question seeing the commercial of Fincad. Who's the king in financial software? Algorithmics??

No it's not. Nobody. Bigger/established groups tend to build their stuff from scratch eg in C++(we certainly do ), canned software is for smaller groups that don't have the quants and devs to do so.

Well, I think there are a lot of sophisticated products and vendors such as RiskMetrics or KMV, Calypso (trading systems) or SunGard (FRONT-ARENA). And a lot of big banks use it! And again, e.g. RiskMetrics is far from being trivial ...

Zmei works for a large bank, so of course they write a higher % of their own stuff."The industry standard" is probably Excel, followed by Matlab, but measured by users I suspect FinCad is not pushing the truth too hard.

Financial software sounds a bit broad and vague.For pricing of derivatives and risk, Numerix pretty well known. Like DCFC says, if your organization is big enough for it to make sense, people prefer to develop their own systems. Big enough to make sense is not even all that big.DCFC: Are you saying that there are smaller banks that actually use Excel to keep track of their derivatives pricing and risk ? Sounds very shoddy.

Excel (or C++) is just a tool. Anything can be harmful is not used in an appropriate way...FinCAD is for example quite common in hedge fund. You won't rely on it for something which is business critical (depending on the nature of your fund that can be LIBOR rate model for those who trades IRS, some form of copula model for credit funds or whatever fun thing you can dream of for algorithmic trading people). But there are asset classes outside your main area that you may need to evaluate from time to time (e.g for benchmarking, marketing etc). Instead of writing something from scratch, it makes sense to use FinCAD in those situations.

What about NAG?

NAG is not exactly financial software We use NAG for all serious numerical work in our C++ code. Its C interface is incredibly painful, but it's super fast and I'm very fond of it. From colleagues who have used both, I've heard that NAG optimizers are superior to both Matlab optimization toolbox and Minuit (the optimizer that comes with root).

Murex is the best in derivatives risk management. Calypso is probably second. Murex provides a solution from back to front office and you can basically price anything except for probably the more complex credit derivatives. Fincad is more appropriate if you are a small hedge fund that wants to get started quickly with the basic stuff.

How about OpenLink

ChicagoGuy, do you really enjoy Murex?Amb

ChicagoGuy, you must be the only living fan of Murex out there... We must put you on some sorta pedestal in some sorta museum. Having 'enjoyed' Mewrecks for 2yrs+ first-hand, I can say that I go into conniptions every time I hear mention of it now. I use FinCAD, however, for quite a few things and it works OK as a a provider of basic analytical capabilities. Once a piece of analysis is fleshed out, it goes on to Matlab. I hear Numerix is a lot better than FinCAD, but I have never had the pleasure.My 2c...

If you use Murex, check every valuation type--e.g. set a double barrier to value a single barrier or even Vanilla; set-up one type of instrument valuation so it can value another type, etc.. I worked with it for a long time [six years ago] and found many obvious valuation errors with just a few minutes of checking and fixes then caused other valuation errors.

FinCAD is pretty lightweight relative to some others mentioned on the forum. (Although pretty cheap too).I agree with ChicagoGuy. Although i'd add ; For the bigger vendors; Calypso is king in the back office. However - in my experience (if the group are not building internally from scratch) if front office have the choice between Calypso and Murex they will usually go for Murex due to existing functions, ability to capture more complex trade types for pricing and finally the flex-API. (I'm not expert in the flex-API) but it has to be better than fiddling with the JNI. That said - Calypso is probably more extensible.

Well, I think that Murex is the most comprehensive in terms of pricing. The downside is that its expensive and it takes a while to implement. As you guys mentioned, it also has bugs. But dont all sophisticated programs have bugs? The newer versions might have fixed some of the bugs.