iskander

68 ms·

661.
by iskander 15y ago
What I'm saying is that high frequency market makers are only need in very small numbers (if even that). When 2/3 of a market's volume is HFT position shuffling, you're not "improving liquidity" in a way that's meaningful or useful to anyon
662.
by iskander 15y ago
>Spread trading (or "market making") is also misunderstood. People see market makers as scalpers when in fact they're providing a valuable service: they're creating liquidity. The reason you can buy or sell shares at any time (rather th
663.
by iskander 15y ago
Re: 10 microsecond latencies. I wonder what a trading algorithm could possibly do in a few thousand clock cycles that is still profitable.
664.
by iskander 15y ago
I use h5py's datasets (which are organized hierarchically and stored in compressed chunks) to do basic filtering and then load a fraction of my data into memory as numpy arrays.
665.
by iskander 15y ago
I'm totally ignorant of mongodb: what does it do for you (in the way of data aggregation) that's not easy in numpy?
666.
by iskander 15y ago
Are there advantages over storing data in HDF? I've been working with a few hundred gigabytes of financial data this summer and I'm finding that python's data-oriented libraries (h5py, numpy, scipy, matplotlib, scikits.learn) cover my needs
667.
by iskander 15y ago
You're falling way below the usual standards of discourse on this site.
668.
by iskander 15y ago
Interestingly, google.hk is blocked everywhere (which wasn't true when I visited last summer).
669.
by iskander 15y ago
Everyone has their own balance. I sold my car 4 years ago (biking feels better), I never bought a TV, etc... On the other hand, I like inviting people over for drinks, potlucks, etc... So, owning decent cookware and some living room furnitu
670.
by iskander 15y ago
For some of us, it will look more like a martingale :-).
671.
by iskander 15y ago
I learned the above by speaking with employees at a private for-profit school, but when I went looking for supporting evidence I found I've been somewhat misinformed. The private school loan default rates are much worse than public schools:
672.
by iskander 15y ago
>The most likely crime being investigated is student aid fraud, which is an immense problem especially at inner-city colleges. Because of the ease of obtaining government grants and loans, students can register for different programs un
673.
by iskander 15y ago
(I'm going to agree with your first sentence by disagreeing with everything you said about market making) > Look at market making (providing liquidity for other traders). There's an intrinsic demand for liquidity from funds which try t
674.
by iskander 15y ago
I read/watch AJ English most days and I've found that their on-the-ground reporting is excellent but they also publish a lot of marginal rubbish.
675.
by iskander 15y ago
I tried about a 100 password/login pairs and none worked. Perhaps they've all been changed, or maybe this list is fake.
676.
by iskander 15y ago
>Probability is not the sort of math this guy is aiming at. I get that, I'm just not sure there's a good reason for it. Extending discrete models to their probabilistic versions is often very useful. Probabilistic knowledge representati
677.
by iskander 15y ago
Strangely, not one word about probability.
678.
by iskander 15y ago
Decision trees suffer from high variance. A slightly different sample might give you entirely different splits; using decision trees for data interpretation or feature selection is an art at best (and for some data sets uselessly unreliable
679.
by iskander 15y ago
http://www.system76.com/product_info.php?cPath=28&produc...
680.
by iskander 15y ago
Admittedly I only programmed in Ada for a few months, but I found it to be an unbearably tedious language. The verbosity is monstrous--- the type system is inexpressive; it feels primitive. Nothing is inferred, everything is repeated. Ugh,
681.
by iskander 15y ago
Ahh, I see, so languages "as given" are those with poor support for metaprogramming? Where would this sort of distinction place Template Haskell, MetaOcaml, or even C++ templates?
682.
by iskander 15y ago
Aren't languages always defined in terms of some semantics or execution model? I guess I don't see why Haskell is more "given" than CL.
683.
by iskander 15y ago
1) I have no idea what is meant here by "given" vs. "medium". 2) In my experience, the difference between C, CL, Forth and Haskell is very small compared to their distance from Prolog. 3) I wonder where APL or Smalltalk are supposed to fall
684.
by iskander 15y ago
You could, you know, spend it on people other than yourself.
685.
by iskander 15y ago
Are there any high profile examples of people who didn't change their lifestyle much at all?
686.
by iskander 15y ago
The few times I've made small utilities in OCaml have proven satisfying (quick to code up, correct once I made it past the type checker). I wonder if the Jane Street folks ever use OCaml in the places people usually stick python or shell sc
687.
by iskander 15y ago
...or you could put a lightly padded mat on the floor. The cost is minimal and it makes my back feel healthier than a bed. As far as I can tell, beds are for comfort not health.
688.
by iskander 15y ago
>Turning to domestic affairs, I wouldn't call China's policies nuanced either. Maybe draconian? Their one-child-per-family policy, their restrictions on speech and travel, their widespread confiscation and displacement, (I could go on)
689.
by iskander 15y ago
I'm an NYU grad student and I think putting city money into my school would be a greater waste. NYU is a distracted and sprawling bureaucracy. Getting two weeks of summer funding required a long series of confusing email exchanges which fi
690.
by iskander 15y ago
Anyone know other examples of senses messing with each other?

More ›