Notices

Curve-fit

From Traderpedia

(Difference between revisions)
Revision as of 13:52, 11 October 2005 (edit)
Rhody Trader (Talk | contribs)
m
← Previous diff
Current revision (13:52, 11 October 2005) (edit) (undo)
Rhody Trader (Talk | contribs)
m
 
Line 1: Line 1:
-{{Glossary|Over-optimizing trading system parameters such that they produce the best possible performance using [[historical data]], but not necessarily going forward.}}+{{Glossary|Over-optimizing [[trading system]] parameters such that they produce the best possible performance using [[historical data]], but not necessarily going forward.}}
In computerised systematic trading, you can optimise trading rules to precisely buy and sell at the most profitable [[swing high]] and [[swing low|low]] for all past market data in your sample. However, if you use TOO much math and optimise your system too precisely to a specific sample of past events (read market prices), then it will almost certainly mean that the strategy won't continue to perform in the future since the events the system was optimized for will not occur in the same way again. In computerised systematic trading, you can optimise trading rules to precisely buy and sell at the most profitable [[swing high]] and [[swing low|low]] for all past market data in your sample. However, if you use TOO much math and optimise your system too precisely to a specific sample of past events (read market prices), then it will almost certainly mean that the strategy won't continue to perform in the future since the events the system was optimized for will not occur in the same way again.
[[Category:terms]] [[Category:terms]]

Current revision

Definition:
Over-optimizing trading system parameters such that they produce the best possible performance using historical data, but not necessarily going forward.


In computerised systematic trading, you can optimise trading rules to precisely buy and sell at the most profitable swing high and low for all past market data in your sample. However, if you use TOO much math and optimise your system too precisely to a specific sample of past events (read market prices), then it will almost certainly mean that the strategy won't continue to perform in the future since the events the system was optimized for will not occur in the same way again.