> In particular, the daily closing price is what is used to determine the buying/selling price when orders are placed on a given day, and orders may only be pla

> In particular, the daily closing price is what is used to determine the buying/selling price when orders are placed on a given day, and orders may only be placed once per day. Umm....... I don't want to shit on someone's hard work and maybe I'm reading this wrong but as an FYI to anyone who wants to try modelling like this.... I read this paper a couple of days ago when someone forwarded it to me but I'll say that any time a model works by buying/selling at the closing price its usually broken. You generally can't just buy at the closing price as you dont' know what the closing price is until after it happens. MOC order accepting markets aside. Or put another way, if you back test your model by assuming you can trade at the closing price its really no different than saying "hey my model works if I can look into the future to see what prices will do and then go back in time and trade with this knowledge". Having said that, here's what I like about this paper.... It goes into atleast a tiny bit of depth covering the mathematics behind their idea. I see lots of programmers who say "I don't really know math but I want to get into algorithmic trading but I don't really know math well, how should I start?" And I think I break more than a few hearts when I tell them to start by learning math. And even if you know math well, I'll leave this for thought..... The biggest mistake that most engineers make is that they figure they'll throw some machine learning at the market and start making money. its my experience that creating a trading strategy is 70% to 90% market intuition with the rest being the math and implementation( where implementation will be the computer code to execute the trade when the conditions are right). Remember, every day 100's of MIT Phd's try to apply machine learning algos to all sorts of markets and instruments. It takes a large amount of hubris to think that someone with a compsci degree and no market intuition can walk in, fire up R and start making money. I don't have a Phd, I'm just a computer engineer who still leans very heavily on my math text's and wolfram alpha whenever I try to solve a differential equation so it can be done, you just need to put in the effort and be willing to lose money while learning!