5 ms·

I don't think they can really draw this conclusion. There are a great number of additional factors at play and accessing public sec records is a bit of a red he

by osrec 9y ago

I don't think they can really draw this conclusion. There are a great number of additional factors at play and accessing public sec records is a bit of a red herring. Especially considering that the funds they have listed as examples are primarily large high frequency algo funds (Renaissance and AQR) and are even less likely to find much utility in intermittent public filings. This may however be an indicator of diligence, but trying to link public information to a big increase in performance is somewhat far fetched.

scarletham 9y ago
Tradeworx is HFT. Yes, the premise of this paper is completely absurd. Drawing a line between overnight batch jobs for internal databases and performance is a quantum leap to say the least.
philfrasty 9y ago
Do you know how these algo funds keep a competitive edge for so many years (e.g. Renaissance 30+ years)? I am trying to understand their "kind" of product innovation as in what are the biggest factors they are continuously trying to improve?
- wtvanhest 9y ago
  They improve their model. The factors that allow them to capture alpha change so you cannot set it and forget it. The holy grail of investing would be a quantitative model which automatically changes and consistently outperforms everyone else.
- dsacco 9y ago
  Renaissance Technologies has completely automated the process of signal discovery.[1] They don't hire researchers to manually derive novel insights or trading models from data, and they don't really bother with exclusive sources of data. Instead, they hire researchers to improve methods for automatically processing vast amounts of arbitrary data and extracting profitable trading signals from it. When most funds say they're "quantitative", what they really mean is that they use huge amounts of data to inform fundamentally manual trading strategies (this includes most of the places widely considered to be "top" firms). They develop trading algorithms, and those trading algorithms are often successful. But the algorithms are developed manually and then deployed. Their researchers and engineers actively seek out new sources of data and try to compete on novel sources of untapped information. But the reality of what happens is that they simply drown in the data. They can't clean it or process it nearly fast enough to maintain long term trading strategies, nor can they even begin to find a way to automate the trading strategy extraction. If you're working with hundreds of terabytes of data, you cannot selectively formulate hypotheses and test them. It's far too slow. You will find dramatically fewer novel insights than a fully automated process. In other words, they're a step above traditional "fundamental" hedge funds, but they focus on the wrong problem (but not for lack of trying!). In contrast, the truly successful quant funds have automated the data processing and feature extraction pipeline end to end. The data is a pure abstraction to them. They don't bother with forming hypotheses and trying to find data to test them, they allow their algorithms to actively discover new correlations from the ground up. So many quantitative funds advertise how much data they work with, and how they have all these exotic sources of data at their disposal...but the data does not matter. The models for the data do not matter. The mathematics of efficiently processing that data are what matters. As a result of their consistent profitability, most of the jobs you see listed for the really successful funds (if they have a website) are not "real" in the strictest sense of the word. You can apply to them, but they only keep active careers pages to attract the best researchers. Their only incentive to hire is to 1) keep someone who is actually exceptional from joining a competitor or 2) keep an academic researcher from re-discovering their work when they seems like they're getting close to it. This is why they primarily focus on quantitative PhDs in information theory, high energy physics and computational mathematics (especially information geometry). To be completely frank, Renaissance is an outlier, but not just because of their returns. They're an outlier because of how public they are. Most of the funds with comparable returns not only don't take any outside investor capital, they only have 25 - 50 employees. They virtually never hire because they don't have to. If your work is fundamentally interesting, novel and applicable to what they're doing (even if you can't immediately see why), they will call you. ___________________________ 1. Other more secretive (but equally successful) funds have done this, but they are much more under the radar.
  - zechs 9y ago
    They automated signal discovery in what kind of data? I have heard their use of unconventional data sources is the source of their success, not automated signal discovery.
    - dsacco 9y ago
      No, every quantitative firm uses unconventional sources of data. That doesn't meaningfully differentiate them (at least, not anymore). For example, Two Sigma has an entire division devoted to sourcing and processing "alternative data." But Two Sigma is not at all comparable to firms like RenTec. The funds I'm talking about (including RenTec) take in as much unstructured data as they can possibly find, almost indiscriminately, and they tune their processing pipeline to the point that it requires neither manual classification nor munging. In most cases, a trading strategy is sufficiently multidimensional that any particular set of data can be completely public. Exclusive data is helpful, but not required. In many cases people become too dependent on exclusive data and lose sight of the methodology. This is precisely what I mean: many people think that these firms differentiate based on the sources of data they use. They do not. They differentiate on their ability to automatically extract signals hiding in plain sight. Whether or not the data is public makes very little difference, because the signals come from tens of thousands of indicators combined together.
      - autokatalyst 9y ago
        You mentioned these firms are typically smaller in headcount, use private capital, and generate similar returns. Any insight on how they're targeting asset capacity for their meta-strategy approach? The medallion fund, until recently was capped around 3-4 billion [1]. I wonder, if the quality and quantity of data itself, and the models built on that data, are subordinate to the data processing IP as you say, can this type of trading be done at smaller scale, either as an individual or handful of people? I've been out of the industry for a bit, but infrastructure and data costs were the main (financial) hurdles for startups, particularly in the high frequency space. Are there any firms outside of finance using an analogous approach to data processing and signal construction? [1] https://www.bloomberg.com/news/articles/2017-08-16/renaissance-s-medallion-made-stunning-shift-after-trump-election https://www.bloomberg.com/news/articles/2017-08-16/renaissan...
jeffreyrogers 9y ago
AQR is algorithmic but not high frequency, as far as I know. No one knows what strategies Renaissance employs.
avvt4avaw 9y ago
AQR are certainly not a high frequency trading firm, and Renaissance may do some high frequency trading, but they are not primarily a high frequency trading firm. Where did you get that idea? The paper explicitly addresses the point that large-scale systematic collection of public records may be indicative of the kind of fund that outperforms, rather than an indication that the public records add alpha in and of themselves - it's in the abstract > The effect is not due to differences in fund type as the results hold within-fund. I think you need a more convincing argument than just saying "There are a great number of additional factors at play" and "Trying to link public information to an increase in performance is far fetched".
- scarletham 9y ago
  They're not HFT, although they are more blackbox / quant than the "traditional" hedge fund, which I think still supports OP's point.
  - dsacco 9y ago
    Most "black box" quantitative hedge funds still use massive amounts of public data, even public data that might appear somewhat obvious. They absolutely develop new insights from that data.
    - scarletham 9y ago
      Yeah, to clarify I was agreeing that those firms certainly pull and use that data, but not necessarily in the same traditional fashion that paper seems to imply.
- osrec 9y ago
  Having worked with AQR, I would say they are more HF + coupled with some tend following than anything else. What makes you say they are not?
  - auntienomen 9y ago
    Does that just mean that they do their own execution, or does it mean that their short-holding-time strategies are a major profit center?

by osrec 9y ago

scarletham 9y ago

Tradeworx is HFT. Yes, the premise of this paper is completely absurd. Drawing a line between overnight batch jobs for internal databases and performance is a quantum leap to say the least.

philfrasty 9y ago

Do you know how these algo funds keep a competitive edge for so many years (e.g. Renaissance 30+ years)? I am trying to understand their "kind" of product innovation as in what are the biggest factors they are continuously trying to improve?

wtvanhest 9y ago
They improve their model. The factors that allow them to capture alpha change so you cannot set it and forget it. The holy grail of investing would be a quantitative model which automatically changes and consistently outperforms everyone else.
dsacco 9y ago
Renaissance Technologies has completely automated the process of signal discovery.[1] They don't hire researchers to manually derive novel insights or trading models from data, and they don't really bother with exclusive sources of data. Instead, they hire researchers to improve methods for automatically processing vast amounts of arbitrary data and extracting profitable trading signals from it. When most funds say they're "quantitative", what they really mean is that they use huge amounts of data to inform fundamentally manual trading strategies (this includes most of the places widely considered to be "top" firms). They develop trading algorithms, and those trading algorithms are often successful. But the algorithms are developed manually and then deployed. Their researchers and engineers actively seek out new sources of data and try to compete on novel sources of untapped information. But the reality of what happens is that they simply drown in the data. They can't clean it or process it nearly fast enough to maintain long term trading strategies, nor can they even begin to find a way to automate the trading strategy extraction. If you're working with hundreds of terabytes of data, you cannot selectively formulate hypotheses and test them. It's far too slow. You will find dramatically fewer novel insights than a fully automated process. In other words, they're a step above traditional "fundamental" hedge funds, but they focus on the wrong problem (but not for lack of trying!). In contrast, the truly successful quant funds have automated the data processing and feature extraction pipeline end to end. The data is a pure abstraction to them. They don't bother with forming hypotheses and trying to find data to test them, they allow their algorithms to actively discover new correlations from the ground up. So many quantitative funds advertise how much data they work with, and how they have all these exotic sources of data at their disposal...but the data does not matter. The models for the data do not matter. The mathematics of efficiently processing that data are what matters. As a result of their consistent profitability, most of the jobs you see listed for the really successful funds (if they have a website) are not "real" in the strictest sense of the word. You can apply to them, but they only keep active careers pages to attract the best researchers. Their only incentive to hire is to 1) keep someone who is actually exceptional from joining a competitor or 2) keep an academic researcher from re-discovering their work when they seems like they're getting close to it. This is why they primarily focus on quantitative PhDs in information theory, high energy physics and computational mathematics (especially information geometry). To be completely frank, Renaissance is an outlier, but not just because of their returns. They're an outlier because of how public they are. Most of the funds with comparable returns not only don't take any outside investor capital, they only have 25 - 50 employees. They virtually never hire because they don't have to. If your work is fundamentally interesting, novel and applicable to what they're doing (even if you can't immediately see why), they will call you. ___________________________ 1. Other more secretive (but equally successful) funds have done this, but they are much more under the radar.
- zechs 9y ago
  They automated signal discovery in what kind of data? I have heard their use of unconventional data sources is the source of their success, not automated signal discovery.
  - dsacco 9y ago
    No, every quantitative firm uses unconventional sources of data. That doesn't meaningfully differentiate them (at least, not anymore). For example, Two Sigma has an entire division devoted to sourcing and processing "alternative data." But Two Sigma is not at all comparable to firms like RenTec. The funds I'm talking about (including RenTec) take in as much unstructured data as they can possibly find, almost indiscriminately, and they tune their processing pipeline to the point that it requires neither manual classification nor munging. In most cases, a trading strategy is sufficiently multidimensional that any particular set of data can be completely public. Exclusive data is helpful, but not required. In many cases people become too dependent on exclusive data and lose sight of the methodology. This is precisely what I mean: many people think that these firms differentiate based on the sources of data they use. They do not. They differentiate on their ability to automatically extract signals hiding in plain sight. Whether or not the data is public makes very little difference, because the signals come from tens of thousands of indicators combined together.
    - autokatalyst 9y ago
      You mentioned these firms are typically smaller in headcount, use private capital, and generate similar returns. Any insight on how they're targeting asset capacity for their meta-strategy approach? The medallion fund, until recently was capped around 3-4 billion [1]. I wonder, if the quality and quantity of data itself, and the models built on that data, are subordinate to the data processing IP as you say, can this type of trading be done at smaller scale, either as an individual or handful of people? I've been out of the industry for a bit, but infrastructure and data costs were the main (financial) hurdles for startups, particularly in the high frequency space. Are there any firms outside of finance using an analogous approach to data processing and signal construction? [1] https://www.bloomberg.com/news/articles/2017-08-16/renaissance-s-medallion-made-stunning-shift-after-trump-election https://www.bloomberg.com/news/articles/2017-08-16/renaissan...

jeffreyrogers 9y ago

AQR is algorithmic but not high frequency, as far as I know. No one knows what strategies Renaissance employs.

avvt4avaw 9y ago

AQR are certainly not a high frequency trading firm, and Renaissance may do some high frequency trading, but they are not primarily a high frequency trading firm. Where did you get that idea? The paper explicitly addresses the point that large-scale systematic collection of public records may be indicative of the kind of fund that outperforms, rather than an indication that the public records add alpha in and of themselves - it's in the abstract > The effect is not due to differences in fund type as the results hold within-fund. I think you need a more convincing argument than just saying "There are a great number of additional factors at play" and "Trying to link public information to an increase in performance is far fetched".

scarletham 9y ago
They're not HFT, although they are more blackbox / quant than the "traditional" hedge fund, which I think still supports OP's point.
- dsacco 9y ago
  Most "black box" quantitative hedge funds still use massive amounts of public data, even public data that might appear somewhat obvious. They absolutely develop new insights from that data.
  - scarletham 9y ago
    Yeah, to clarify I was agreeing that those firms certainly pull and use that data, but not necessarily in the same traditional fashion that paper seems to imply.
osrec 9y ago
Having worked with AQR, I would say they are more HF + coupled with some tend following than anything else. What makes you say they are not?
- auntienomen 9y ago
  Does that just mean that they do their own execution, or does it mean that their short-holding-time strategies are a major profit center?