Comments on: The Relevance of History http://keplerianfinance.com/2013/06/the-relevance-of-history/ exploring the boundaries of quantitative finance Wed, 14 Aug 2013 02:24:36 +0000 hourly 1 https://wordpress.org/?v=4.1.41 By: Robert J Frey http://keplerianfinance.com/2013/06/the-relevance-of-history/#comment-55 Tue, 16 Jul 2013 07:28:18 +0000 http://keplerianfinance.com/?p=277#comment-55 Thank you for a careful reading. I’ve fixed the typo.

]]> By: David Haynes http://keplerianfinance.com/2013/06/the-relevance-of-history/#comment-52 Fri, 12 Jul 2013 02:55:30 +0000 http://keplerianfinance.com/?p=277#comment-52 There’s a typo for the annual standard deviation formula. The easiest way to fix it would be to remove the monthly standard deviation from under the square root sign.

]]> By: clash of clans fan http://keplerianfinance.com/2013/06/the-relevance-of-history/#comment-51 Wed, 10 Jul 2013 17:24:27 +0000 http://keplerianfinance.com/?p=277#comment-51 clash of clans fan…

Very nice post. I just stumbled upon your blog and wished to say that I have truly enjoyed browsing your blog posts. In any case I will be subscribing to your rss feed and I hope you write again very soon!…

]]> By: MOB http://keplerianfinance.com/2013/06/the-relevance-of-history/#comment-47 Wed, 03 Jul 2013 16:01:00 +0000 http://keplerianfinance.com/?p=277#comment-47 Prediction requires a more precise answer rather than a probability of occurences. This is more of dictation and interpretation of results based on non traditional mehodologies. Here is where the highly complex concept of alogorthms come into play. There has to be theory formed to create a singular result and a table to back it up. Happy Hunting!!!!!!

]]> By: S http://keplerianfinance.com/2013/06/the-relevance-of-history/#comment-44 Sat, 29 Jun 2013 06:39:43 +0000 http://keplerianfinance.com/?p=277#comment-44 Excellent post. Much enjoyed and appreciated.

]]> By: Algomind http://keplerianfinance.com/2013/06/the-relevance-of-history/#comment-43 Fri, 28 Jun 2013 11:08:17 +0000 http://keplerianfinance.com/?p=277#comment-43 >I expressed my concern that data at that frequency over such a limited period probably weren’t enough

You were perfectly right, imho. I’d say more. There is no amount of past data which will ever “suffice” in that direction.

Why ? A a trading methodology must be self-consistent and cannot rely on a process of selection based on past data (“curve fitting”), as this would actually violate all the conceptual basis of Statistics. (I won’t go in deeper details, as this would take way too much space)

A more meaningful way to “explore” methodologies is through extensive simulation of future scenarios (eg, random tickdata based on process mixtures), that is scanning the entire sample space. But it is, in any case, more a tool and a process to refine the trading concepts, than a way to select or predict anything.

]]>