This question struck me a couple of nights ago while I was working on some homework. Since then I haven’t been able to get it out of my head. In linear algebra we often use i and j. a b and c get come into parlance as soon as geometry. And x and y are in just about every single equation.
But what about g? Or s? A person far more gifted in mathematics than I suggested o as a possible answer. I’d like to get the straight dope’s thoughts on this question as well.
Probably e, because it almost represents the mathematical constant e (2.71828…), not a variable.
g and s are commonly used in physics and engineering. (And of course the integral sign is just a stretchy S.)
I would guess o. It’s too easy to confuse with zero.
Edit: Of course, capital O is ubiquitous in computer science, so…yeah
or l, which is easy to confuse with 1 or I. Whenever I make the mistake of using l in a program as a variable, I regret it.
Eh, I’ve seen people use both e and pi as variables before, probably just to be annoying. There is also the famous pi function which is related to the Riemann hypothesis.
o (and O) are common enough. (Maybe that is not enough of a “variable”?) So is g. S occurs most often as the sign for an integral (definitely not a “variable” in that case) but that is hardly the only common use. I would be surprised if any Latin letter is especially rare, though. I would wildly guess the frequencies only really start dropping maybe with certain Greek letters (lower-case υ ??) and unquestionably plummeting once you get to Hebrew and Cyrillic.
Yeah, that would probably be my first guess, although e (most commonly used to represent the constant 2.718…) and f (most commonly used to represent a function—unless you count that as a variable) are good guesses too.
And that O isn’t being used as a variable, is it?
No, it’s used as a function. So if you want to consider it a variable depends on whether or not you like functional programming. 
If e is used as a variable, I guarantee you that there are also variables a through d. It’s almost always a constant, either the base of natural logarithms, or in physics the charge on an electron. Lower-case L is fine if you write it in script, which it always is by anyone with any training in writing mathematics, and is extremely common in quantum mechanics. F and g are often the names of functions and not as often the names of non-function variables, but functions can definitely be a variable and so there are definitely many times where f and g get used as the names of functions that are variable/unknown as opposed to constant/given.
I would venture a guess that o is the least used because it’s the only one I cannot think of as having used as a variable at any time in my life. Big-O notation and variants are often used in Math/CS/Physics, but it’s not a variable there.
If you extend the meaning of “alphabet”, I bet the answer is “thorn” or similar.
In abstract algebraic stuff e is used a lot. In particular you can see stuff with e[sub]n[/sub] type subscripting. A lot of times there’s elements labeled a, b, c, d, e, etc. The problem comes when you get to i and you’re talking about things similar to imaginary numbers. Oops.
My first thought was o as well. (Not to be confused with ∘ which is used as a generic group operator, function composition, etc. and not as a variable.)
No love for j?
j is commonly used in engineering in place of i in imaginary numbers, since i is commonly used for current (from the french intensitie).
The three simplest and most basic components in electricity are the resistor, inductor (a coil of wire) and a capacitor (two metal plates). The impedance of a resistor is R. The impedance of an inductor and capacitor depend on the frequency, and are jwL and 1/jwC respectively, where j is actually i for complex numbers and w is actually the greek omega, but is often typed as a w just since it is quick, easy, and close enough that everyone knows what you’re talking about.
If writing it out by hand we’ll use the real omega instead of w, but we still use j just so we don’t confuse it with i for current.
So yeah, j gets used a LOT in engineering math.
I’ve seen it as an index on tensors where there are four of them. It’s usually italicised or written in cursive to distinguish from “1”.
But that would make it a constant, not a variable.
I believe i and j are both used fairly frequently as index variables, in math and computer programming.
I think that the reason this question is so tricky is because few people take all of the upper level mathematics courses. Whether you approach the subject from an academic, applied, CS, or engineering perspective definitely changes your intuitive answer.
Until I took linear algebra I definitely would have guessed i or j. So far it sounds like o and e are the most likely contenders. The only usage for o I can think of is Big O, but it’s not really being used as a variable in that context. And e would be pretty much off limits due to confusion. I’m definitely interested in hearing more answers though!
As I think about it, I get into trouble when I’m using lower-case L in script for angular momentum, and then transfer my equation to a computer where lower-case L is much more confusing
In upper-level mathematics, e is commonly used to denote the identity element of a group.
In my experience doing theoretical physics, o (particularly the lower-case o) is the least frequently used Roman letter. Even e gets used more often, typically in contexts where exponential functions aren’t really used and the author wants to define a series of variables a, b, c, … Capital O is really only used in the context of big-O notation, and even then it’s commonly written in a special calligraphic font to avoid confusion.
Also, the confusion between “1” and lower-case “l” isn’t as much of an issue as you might think. In well-typeset mathematics, variables are usually in italics while numerals are upright. So the numeral is “1” and the variable is “l”.
In any context where you’ve got points labeled by letters, you’re likely to get one labeled “o” or “O” (even if you haven’t gotten that far in the alphabet), for the origin. Of course, this is one context where it won’t really matter if the o gets mistaken for 0. Other than that, the only places I can think of seeing it are as a label, in some rare context where 15 or more things need labeling.
Wait, I take that back: When teaching trigonometry, one sometimes sees o used for the side of a right triangle opposite of a particular angle (as in soh-cah-toa). But that’s mostly only used in teaching: Right triangles encountered “in the wild”, so to speak, will usually have some other labeling.
If we’re getting into the Greek alphabet, lower-case zeta and xi are very seldom used, mostly because they’re difficult to draw correctly. But I’ve seen them show up as indices in general relativity, occasionally. And I like to use them, because I’m a bit proud of how well I can draw them.
Zeta and xi are used a lot, especially in number theory as functions (often with s as a variable argument). They occur as variables as well, quite often enough that I feel “very seldom used” is incorrect.
I don’t recall Xi being used very often in number theory, but that class was a long time ago for me. I do remember when it was used that it annoyed me, because it was hard to draw.
I think Aleph, in set theory, is the only character from a script other than Greek or Latin to be used.
