Good questions. M
---------- Forwarded message ---------- From: Paul Nash Date: Friday, April 19, 2013 Subject: Re The risks of/when not releasing your code & data: Shocking Paper Claims That Microsoft Excel Coding Error Is Behind The Reinhart-Rogoff Study On Debt To: [email protected] Quite apart from being "clumsy" with their Excel model, they forgot the first rule of research: correlation does not imply causation. So when are they going to resign, and when are the various central bankers who used their model to impose austerity going to change tack? Or will they just brush it aside and get on with screwing the working man? paul On 2013-04-17, at 8:35 AM, Dave Farber <[email protected]> wrote: ---------- Forwarded message ---------- From: Bob Frankston Date: Wednesday, April 17, 2013 Subject: The risks of/when not releasing your code & data: Shocking Paper Claims That Microsoft Excel Coding Error Is Behind The Reinhart-Rogoff Study On Debt To: [email protected], ip <[email protected]> Obviously I've had some experience with spreadsheet "errors". Of course they are just tools. The deeper issue is long chains of reasoning without reality checks via some other approach. It doesn't matter if it's a "coding" error or a reasoning error. How do you do a reality check on theories and, more to the point, predictions. What are the system assumptions. I've cited Black-Scholes numbers with Interactive Data used to provide "values" for options in portfolios in the early 1970's. The purpose was to give people a number that made them happy. Unfortunately some people thought those numbers had predictive value. In general beware of any results that "prove" what the modeler wants to prove. Science proceeds by testing results not simply confirming assumptions. This reduces the harm done by even "correct" models in changing contexts. From: Dave Farber [mailto:[email protected]] Sent: Wednesday, April 17, 2013 06:52 To: ip Subject: [IP] The risks of/when not releasing your code & data: Shocking Paper Claims That Microsoft Excel Coding Error Is Behind The Reinhart-Rogoff Study On Debt ---------- Forwarded message ---------- From: the keyboard of geoff goodfellow Date: Wednesday, April 17, 2013 Subject: the risks of/when not releasing your code & data: Shocking Paper Claims That Microsoft Excel Coding Error Is Behind The Reinhart-Rogoff Study On Debt To: Dave Farber <[email protected]>, ip <[email protected]>, Dewayne Hendricks <[email protected]>, [email protected] see it all for yourself: http://www.businessinsider.com/thomas-herndon-michael-ash-and-robert-pollin- on-reinhart-and-rogoff-2013-4 The conclusion: ... This error is needed to get the results they published, and it would go a long way to explaining why it has been impossible for others to replicate these results. If this error turns out to be an actual mistake Reinhart-Rogoff made, well, all I can hope is that future historians note that one of the core empirical points providing the intellectual foundation for the global move to austerity in the early 2010s was based on someone accid <https://www.listbox.com/member/archive/247/=now> Archives <https://www.listbox.com/member/archive/rss/247/22720195-c2c7cbd3> | <https://www.listbox.com/member/?member_id=22720195&id_secret=22720195-8fdd4 308> Modify Your Subscription | <https://www.listbox.com/unsubscribe/?member_id=22720195&id_secret=22720195- 97c5b007&post_id=20130419121834:C6A7FD72-A90C-11E2-86BC-A79BB01A6B3F> Unsubscribe Now <http://www.listbox.com>
_______________________________________________ Futurework mailing list [email protected] https://lists.uwaterloo.ca/mailman/listinfo/futurework
