In technical analysis, we may use confluence of direction for 3 timeframes to roughly gauge bias of market now. Similarly, if we use time series forecasting methods to predict(say daily data-whether S&P is going higher tomorrow), how much historical daily data would be optimal(bet 2 weeks-1month-3 months)? Too much or too little past data does not give accurate prediction.

(1) generate results in 5 days intervals(within 3 months) until you get the best interval prediction that is closest to yesterday's closing value...Then use this interval for predicting tomorrow's close?

(2) combine 3 months forecasting and backcasting(reverse data) until there is a result that coincides...then use this day as starting reference point for forecasting? http://www.spiderfinancial.com/support/documentation/numxl/tips-and-tricks/backward-forecast http://pakaccountants.com/what-is-backcasting-and-difference-forecasting/

Other suggestions?

  • 1
    $\begingroup$ Careful! You may become the victim of overfitting if you search for minor variables like this. $\endgroup$ Commented Jan 15, 2013 at 12:14

1 Answer 1


You should never start out asking how much data you should incorporate in your research effort. You should start with the following points:

  • Make sure you understand the difference between sampling size to specify your model and data used to run back tests over. Those are entirely different animals.
  • First, think about your end goal, what are you trying to achieve. Do you look to develop a high frequency trading model? If yes it makes no sense to incorporate price intelligence from 1 month ago.
  • What dynamics are you trying to capture? If you look to trade market cycles then you want to incorporate the amount of data that covers different market cycles.
  • What type of data are you trying to analyze? Tick, compressed, bid/ask vs. trades, order book dynamics?

When you answer those question you should most likely have a very good idea how much data you need, at what frequency.

  • $\begingroup$ Backtesting and BackCasting are entirely different animals. $\endgroup$
    – Shelagh
    Commented Jan 15, 2013 at 12:42
  • $\begingroup$ never heard of the term "back casting"...and your point being? Just glimpsing over your link it does not sound applicable at all. The market cares very little what future state you think the market should reach. You may have an opinion, whether that is consensus is very hard to impossible to estimate. You just made your modeling and valuation effort twice as hard as it should be $\endgroup$
    – Matt Wolf
    Commented Jan 15, 2013 at 12:48
  • $\begingroup$ that is nothing else than back testing, just that you forecast past prices. I am not sure where you are trying to get with this, your question asked mostly what time frames to incorporate when analyzing time series. I provided a guideline how to arrive at the answer you sought. $\endgroup$
    – Matt Wolf
    Commented Jan 15, 2013 at 12:54
  • $\begingroup$ Could it be where the forecasted past price and forecasted future price coincide closest, the better starting point of reference for forecasting? $\endgroup$
    – Shelagh
    Commented Jan 15, 2013 at 12:59
  • $\begingroup$ well that is the whole idea why people back test. Such models forecast the "future past", then the model is calibrated so that the forecast prices match up with realized prices and the calibrated model is used to then forecast truly future prices. Attention has to be paid to not overfit such models. $\endgroup$
    – Matt Wolf
    Commented Jan 15, 2013 at 13:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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