Hot answers tagged performance
19
Column-oriented storage is faster for reading because of the cache efficiency. Looking at your sample query:
select price, time from data where symbol = `AAPL
Here I'm concerned with three columns: price, time, and symbol. If all ticks were stored by row, the database would have to read through all rows just to search for the symbols. It would look like ...
11
I have long hungered for the ultimate, super-fast, super-scaleable data storage solution. I have used relational databases, kdb, flatfiles, and binary files. In the end, I used binary files in my research language of choice. My advice is to KISS. The choice of storage is actually not that critical (unless maybe you're working with options tick data). ...
8
Personally I make a distinction between two conflicting goals: (1) storing data incoming in real-time for immediate processing and (2) storing the gathered data for "offline" purposes. Such approach makes things a lot easier if we're talking about a home-grown solution.
(1) must be as fast as possible but not necessarily scalable beyond a few dozen millions ...
7
Yes. First, it is much easier to proceed if you standardize the output of your forecast so they are in the same units (returns, for example, or probabilities of an event/condition occurring). After you have done this, there are 3 general approaches:
Signal weighting: Then you need to define a weighting scheme for your factors. Richard Grinold has an one ...
6
Coming from an HPC background myself, I know too well the feeling of owning a hammer and yet having no nail. Your question is about computational bottlenecks that can be relieved with GPGPU, though I'm afraid to admit that there aren't many in finance. For realtime applications, the network is the bottleneck; for historical applications, the memory is the ...
6
An index is just an abstract concept and does not hold securities. Hence no source of revenue from lending them.
A portfolio mirroring an index holds the securities and can in fact generate revenue by loaning the securities to others wanting to short the stocks. This provides a positive bias. That is often offset by a negative bias when the index ...
6
This depends a little bit on your definition of volatility arbitrage but in general what is meant is a strategy that takes advantage of the difference between implied volatility and realized volatility. Normally you receive implied variance and pay realized variance.
This strategy is the classical example of picking up nickles in front of a steamroller ...
5
The answer your are looking for might be the story in "Benchmarking Measures of Investment Performance with Perfect-Foresight and Bankrupt Asset Allocation Strategies", by Grauer (Journal of Portfolio Management).
While this work main concerns are the differential ranking of various performance measures and with negative betas for market timing strategies, ...
5
Your example shows a fundamental ignorance of how hedge funds operate:
Hedge funds cannot advertise and are limited to 499 investors. Given these restrictions plus the capital requirements to hold positions overnight, it is a virtual guarantee that a fund would not take an investment of $10K.
Hedge funds are usually LPs, which means that the GP (the asset ...
5
Here are couple references. Especially the first link to Andy Lo's paper contains a list of Sharpe ratios of popular mutual and hedge funds:
The Statistics of Sharpe Ratios
Dow Jones Credit Suisse Hedge Fund Index
Hedge Fund Performance and Generalized Sharpe Ratios
I would go with the first paper.
5
I have been using FastBit for a while now and find it to be quite performant. It's very non-intrusive to your existing binary storage format provided your data is stored in a columnar manner.
I have briefly tested Tokyo/KyotoCabinet and didnt find it suitable for my (persistent storage) requirements.
4
Thanks gappy for your precise response. However the answer to this auto-correlation is much more important than an academic discussion of which portfolio performance ratio is best. Auto-correlation distorts max draw-down calculations raising the question of whether the (positive) auto-correlation will continue in the future producing large draw-downs, or ...
4
Are there any other mechanisms at play here which might explain this kind of tracking error?
Dirk is right, you often lend the titles internally or not, etc.
You can also write calls for your index, this is not orthodox, but it's ETF, there is no orthodoxy there...
Edit : With the graph and given the outperforming is seasonnal (around May), I think we ...
3
Perhaps check out Poti and Levich (2009), or in a different setting but from one of the same authors, Poti and Wang (2010) "The coskewness puzzle" in JBF. They directly address the issue of what level of SR is plausible.
3
Pardon the lack of an actual link, and the formatting, but in footnote 6 of "Alpha is Volatility times IC times Score", Grinold, Richard C.,
Journal of Portfolio Management, Summer 1994 v20 n4 p9(8), Grinold suggests that "a truly outstanding manager" might have an information ratio of 1.33:
(6) A rough guideline for determining the required IC comes from ...
3
I would even stick to the original paper by Sharpe (1966):
Mutual Fund Performance. The Journal of Business Vol. 39, No. 1, Part 2 pp.119--138
If you look at the numbers on Page 6 you can see that the funds sharpe ratios roughly are between $0$ and $1$.
Since the Sharpe ratio already adjusts for the risk-free rate, you cannot really argue about its ...
3
In my mind, there are two questions here:
1) How does DB make money given a zero expense ratio?
This is covered by Dirk and Lliane. Basically, DB gets cheap funding and stock loan fees in return for paying marketing / index / hedging costs. The ETF investor gets zero expense ratio in return for taking DB credit risk.
2) Why does it look like the etf ...
2
I have become a fan of SQLite. It's a very lightweight SQL database, which you can use as an intermediate solution. I agree with Rich C that the best thing to do is probably come up with a custom solution that is optimal for your needs. Using SQLite as persistent storage, and loading the data in memory when you want to do intensive computations on it seems ...
2
In exotics options pricing, there are lots of CPU bottlenecks -- for example the calculation of Fast Fourier Transform or Monte Carlo simulation. When I price a range accrual in Libor Market Model, I don't use a lot of data (carefully optimized, everything should fit in a few MB of L2 cache), but I do a lot of calculations. This is where, I think, a GPU may ...
2
Whatever method you use, I recommend you test your implementation with Monte Carlo simulations as well as real data (although doing the latter subjects you to data mining bias, it can give a sanity check on your Monte Carlo simulations.) For most instances of multiple algorithms, the returns streams will not be independent, and you should take this into ...
1
Practically, the best metric is the one your boss wants you to use.
Alternately, you can think of the return of the desk like a leveraged security, as described here. This would suggest that the daily performance would be calculated as the profit divided by the basis. If you want to express the return in terms of the capital requirement, then that's one ...
1
If you are using the Bisons model the frequency of compounding should be as much as the the rebalance frequency of the portfolio. In your case to get meaningful results you will need to use daily compounding. One this to be careful is that you must take transaction effect into account when calculating the return. If you are using a software like FactSet it ...
1
As you mention neural network, in general, you may like to look further into various machine learning techniques.
On that side, Quant Guy also mentioned ensemble learning which is the general term to combine different learning models. I'd like to elaborate on this point a bit further:
In machine learning, traditional ways to combine models are simple ...
Only top voted, non community-wiki answers of a minimum length are eligible
