Talk:Backtesting

• The quality of this page is miserable. I recommend that it be subsumed in "Cross-validation", a serious statistical methodology.

• In my opinion backtesting is a serious statistical methodology. I think the subject is most worthy of a page. Sirrom8 (talk) 15:59, 10 September 2009 (UTC)

• Agree w/ Sirrom8. Improved the definition to better explain backtesting and its applications. Jgonion (talk) 01:24, 21 December 2009 (UTC)

• I would suggest to add a link toward Cross-validation_(statistics), because back-testing is a special case (but important case) of cross-validation primarily when applied to Time Series data.--Joannes Vermorel (talk) 18:25, 28 November 2011 (UTC)

• I edited the page to make it clear that this is jargon from the financial industry and to remove climatology, since I do not believe the term is used by climatologists. More work needs to be done to integrate it into statistical modeling pages and there is a need for references. I do think that most of the warnings (not relating to references) can be removed at this point.— Preceding unsigned comment added by 66.78.236.243 (talk) 20:33, 20 March 2013 (UTC)

• Just added a citation because backtesting is a popular but flawed aspect of technical analysis and it is important to mention this.-- Gabassett 08:52, 19 May 2013 (UTC)

• Facebook researchers describes what seems to be another kind of backtesting : after doing an experiment, a small number of Facebook users are assigned to the winning and original conditions. AndrewHZ (talk) 19:05, 1 April 2014 (UTC)

• Not sure what the paragraph about Dake Chen has to do with defining the term. It's almost like an "In popular culture" section for a statistical method. I'm new and don't know the protocol for inclusion. — Preceding unsigned comment added by 216.152.103.25 (talk) 19:14, 7 February 2019 (UTC)

Merging Hindcast and Retrodiction

The two pages Hindcast and Retrodiction appears to me as terminology variants of backtesting, merely applied to different fields. As backtesting is the dominant terminology in statistics and more applied sciences, I would suggest to merge those articles into the backtesting entry.--Joannes Vermorel (talk) 08:57, 8 August 2013 (UTC)

Ok about Hindcast (merged!), but please check and vote here about Retrodiction, that is another concept. --Krauss (talk) 17:30, 17 December 2014 (UTC)