Estimating missing data

I'm working on a C++ forex simulator (hard, hard work).

Just wondering if anyone here knows the best algorithm for estimating missing data. eg. If a bar's OHLC values are 21,25,13,19 and theres an OCO order for a buy @ 25 and a sell @ 13 whats the best (most accurate) way of measuring which one is more likely to have been triggered first.

Should I consider the values of previous bar(s)? I can't look into the future, well technically looking at the close price is looking to the future. Should I just use the least value of high-open / open-low

My brain is really fried with this one..

Something to think about:

http://www.google.com/url?sa=t&sourc...GDeAhDmQEfHhjg


Your best bet would be some form of interpolation method. I'm not too good with C++, but maybe if the end-user could select the type of interpolation allowed, that could help determine how much the profitability varies.

nice link...