That depends on what you mean by “business mathematics.” If business mathematics is accounting and book keeping and making sure you know where the money in your business is going, then great! Your goal is totally achievable so in that sense there are no limits.

If your goal is to model the stock market with arbitrary precision, then you have at least four limits I can think of: empirical limits, computational limits, modeling limits, and chaos.

First, empirical limits. If you want to know exactly what a stock is going to do on a given day, you need to know how much of it is being bought and sold at what price. But lots of people buy and sell stock for lots of reasons, and you don’t know exactly how or why they are doing that. Without this knowledge, your model is operating on incomplete information, and all the math in the world won’t shore up that gap.

Now, suppose you did get a perfect read on every single person on the planet. You still need to compute what is happening in the stock market. That’s a whole lot of interacting variables you are trying to deal with, and we are far from having enough computational power to manage it.

Okay, fine. Maybe you don’t need to model every last little detail. Maybe you can use an efficient market hypothesis, some priors about certain industries, sampling of recent market movements and start thinking of the market as its own thing. You are modeling the market while leaving out mostly irrelevant variables, but your guess is only as good as your model, and your model can’t be perfect if it is leaving things out.

And it is worse than that. Even fully determinate systems with only three moving pieces can demonstrate chaos, which means that arbitrarily large qualitative changes in outcomes can result from arbitrarily small changes in inputs. So unless you know everything your model uses with infinite precision, your model still could tell you nonsense.