5
$\begingroup$

In the video James H. Simons: Mathematics, Common Sense and Good Luck, James said that the only rule (in Renaissance Technologies Co.) is too "never override the computer" (See between 49:00 and 49:20). Could anyone be able to explain what it means and what it is the case?

$\endgroup$
  • 2
    $\begingroup$ Can you possibly provide context or summarize what he said here? The video is over an hour long. $\endgroup$ – dsaxton Sep 14 '16 at 1:28
  • $\begingroup$ I have edited the question; the last words of James are between 49:00 and 49:20. (English is not my first language) $\endgroup$ – Sandra Ross Sep 14 '16 at 1:30
  • 1
    $\begingroup$ Sandra, note that you can link to a time index. $\endgroup$ – GeoMatt22 Sep 14 '16 at 2:01
  • $\begingroup$ Context is crucial. Some people rely on software to suggest defaults in their analyses and consequently they suffer by not using their judgment to override them. $\endgroup$ – whuber Sep 14 '16 at 13:05
7
$\begingroup$

Simons is talking about his company, which uses quantitative methods to automate financial trades of various kinds. "Overriding the computer" means making a trade other than what the computers are currently programmed to make, or preventing a computer-scheduled trade. The reason, Simons goes on to say, is that statistically analyzing the company's performance becomes much harder in the presence of human whimsy affecting the company's trades.

This is closely related to the issue of actuarial (quantitative, automatic) versus clinical (qualitative, manual) decision-making that's considered in psychology. Not surprisingly to a statistician, actuarial methods perform better than clinical methods, and purely actuarial methods perform better than actuarial methods with optional clinical overrides. See, for example,

Dawes, R. M., Faust, D., & Meehl, P. E. (1989). Clinical versus actuarial judgment. Science, 243(4899), 1668–1674. doi:10.1126/science.2648573

$\endgroup$
  • 3
    $\begingroup$ Funnily enough, when actuaries talk about 'actuarial judgement' they don't mean what that article (and others like it) mean. They mean something much closer to the professional judgement that the "clinical judgement" of the title intends -- actuaries regularly make the personal judgements that those articles about clinical judgement show the dangers of ... just in different contexts to this one. When they do make those kinds of judgements they're just as subject to the same kinds of biases, with the same bad results. $\endgroup$ – Glen_b Sep 14 '16 at 2:11
  • $\begingroup$ @Glen_b Good to know. I already thought "actuarial" and "clinical" weren't the best choices of term. $\endgroup$ – Kodiologist Sep 14 '16 at 3:11
  • $\begingroup$ @Glen_b I'm actually not totally clear on the vocabulary here. Does 'actuarial judgement' mean the kind of judgement that actuaries use? An alternative interpretation might be like the way "mechanical thinking" could be used, not with reference to how mechanics think, but rather 'in the style of a mechanism', or something. I imagine, for example, mechanics might say that their thinking is of the 'clinical' sort as well, not mechanical thinking. Actuarial judgement could perhaps be interpreted to mean like 'on an actuarial basis' or something that could give a little more technical vibe? $\endgroup$ – Hatshepsut Sep 14 '16 at 6:32
  • $\begingroup$ In the paper title it means "applying actuarial methods" (in a particular specific sense) and is in effect a kind of mechanical-but-statistically-based calculation. In actuarial literature it means something along the lines of "applying the professional judgment of a trained actuary". $\endgroup$ – Glen_b Sep 14 '16 at 10:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.