And these are the factors in the model.

Yeah, they could-- signals.

We think of them as signals.

Like you can think of, again, a really simple one.

We were talking about Fama French and value book to price is a very simple value signal, okay? So some things that we tend to have, again, this is not universal.

Uh, there might be some exceptions with machine learning.

Hey, maybe start off with a, an economic thesis on why this signal should predict returns, right? Is there, oh, some behavioral finance underreaction story, et cetera? So that you're not just sifting through the data trying to find relationships that maybe aren't gonna hold out a sample because you were just finding patterns in randomness.

What kind of questions are you asking yourself these days?

And I was hoping maybe you could talk a little bit about the different dimensions of that choice and the impact it has on the financing rates and the potential risks that go along with it.

Yeah.

So let's use the S&P five hundred as an example because that is such a popular market.

There's so much excess demand for unfunded exposure, i.e., derivative exposure to the S&P.

The spread is actually pretty high in the S&P five hundred.

Like, if you look at a futures contract, it might be seventy, eighty basis points annually.

So that's important to keep in mind because every basis point counts.

And so I'm curious, in your experience, like what's your read on how allocators are thinking about what sorta alpha programs they should be adding, and why they're overwhelmingly choosing something like equity market neutral?