I really like the idea of obscure bar times on intraday charts... Even better if the data is a year or two old. I guess the best approach would be to mix all these ideas together and not just use one. .
And depending on the time frame you choose, you would most likely be merging today's close and the next day's open into one bar which would definitely make it difficult for someone to recognize or reverse engineer the chats. (Besides, I think most people don't have enough time to waste trying this.) I recommend using at least 60m bars or higher to preserve the TA characteristics in the real charts. There is a lot more noise in the lower time frames, and also if you will be merging the close and tomorrow's open in one bar you want a longer time frame so the effect of the overnight gap is minimal.
Where's the volume?? Volume often gives cues that make it much easier to distinguish real charts versus random. Without volume my guess would be: Random Random Real Real Real Random Random Real Random Random
I think it's a valid exercise without volume. Besides, I think volume would be easy to "fake" by varying it to some degree similar to the price change magnitude. e.g. large range days and gap moves would see higher volume, just like in "real" charts. Also, you could create a synthetic instrument where the resulting chart would be "real" but there would be no volume characteristics.
Just an update. I will be posting the answers to the challenge on Monday. So if anyone else wants to participate, you have a couple more days: http://users.rcn.com/jmcgraw/ .
The Results have been posted: http://www.elitetrader.com/vb/showthread.php?s=&postid=169474#post169474
I dont think I really understand your question. But, go here: http://www.elitetrader.com/vb/showthread.php?s=&threadid=11820&perpage=6&pagenumber=1 And go to the top of the thread to see the answers. For the original charts, go here: http://users.rcn.com/jmcgraw/ I hope that helps.