For the second part of last week’s comics, there’s several strips whose authors prefer to use a single name. I’m relieved. Somehow my writing seems easier when I don’t have a long authorial credit to give. I can take writing “Zach Weinersmith” fourteen times a week. It’s all those appearances of, like, “Corey Pandolph and Phil Frank and Joe Troise” (**The Elderberries**) that slow me way up.

Darrin Bell’s **Candorville** for the 4th shows off one of the things statistics can do. Tracking some measurable thing lets one notice patterns. These patterns might signify something important. At the least they can suggest things that deserve more scrutiny. There’s dangers, of course. If you’re measuring something that’s rare, or that naturally fluctuates a lot, you might misinterpret changes. You could suppose the changes represent some big, complicated, and invariably scary pattern that isn’t actually there. You can take steps to avoid how much weight you give to little changes. For example, you could look at running averages. Instead of worrying about how often Lemont has asked for his clippers this year versus last, look at how often he’s asked for it, on average, each of the last three years, compared to the average of the three years before that. Changes in that are more likely to be meaningful. But doing this does mean that a sudden change, or a slight but persistent change, is harder to notice. There are always mistakes to be made, when analyzing data. You have to think about what kinds of mistakes you would rather make, and how likely you want to make them.

C-Dog talks about fitting Lemont’s hair growth to a curve. This means looking at the data one has as points in space. What kinds of curves will come as close as possible to including all those points? It turns out infinitely many curves will, and you can fit a curve to all the data points you have. (Unless you have some inconsistent data, like, in 2017 Lemont asked both 14 times and 18 times.) So to do an interpolation you need to make some suppositions. Suppose that the data is really a straight line, with some noise in it. Or is really a parabola. Really a sine wave. Or, drawing from a set of plausible curves, which of those best fits the data?

The Bézier Curve mentioned here is a family of shapes. They’re named for Pierre Bézier, an engineer with Renault who in the 1950s pioneered the using of these curves. There are infinitely many of them. But they’re nice to work with. You can make great-looking curves as sharply curved or as smoothly curved as you like, using them. Most modern fonts use Bézier Curves to compute the shapes of letters. If you have a drawing program, it’s got some kind of Bézier curve in there. It’s the weird tool with a bunch of little dots, most of which are nowhere near the curve they draw. But moving the dots changes the way the curve looks.

A Bézier curve can be linear; indeed, it can just be a line. C-Dog’s showing off by talking about a linear Bézier curve. Or he means something that looks a lot like a line, to the casual eye. Negative-sloped means what it would in high school algebra when you talk about lines: it’s a thing with a value that decreases as the independent variable increases. Something getting rarer in time, for example.

Samson’s **Dark Side of the Horse** for the 4th is our Roman Numerals joke for the week. The Roman Numerals scheme has well-defined letters to represent the numbers up to 1,000. It doesn’t really have consistent schemes past that. But then the Roman Numeral scheme was a bit more ad hoc than really seems comfortable, to us. There could be a striking variety of ways to write larger numbers, particularly; MathWorld notes how letters like I or X or C would be framed in different ways to get at huge numbers like a hundred thousand or so. Roman Numerals standardized in the middle ages, long after the Roman Empire had reason to care about them, and for that matter as Arabic numerals got to be more accepted. Wikipedia also lists a bunch of Medieval abbreviations in a Roman Numerals scheme for things we just don’t use, like F for 40 or T for 160. I presume they have abundant manuscript examples of these, so that we aren’t making too much out of one person’s idiosyncratic notes.

Thaves’s **Frank and Ernest** for the 5th uses arithmetic, particularly simple addition, as emblematic of the basics of life. Hard to argue that this isn’t some of the first things anyone would learn, and that mathematics as it’s taught builds from that. A mathematician might see other fields — particularly set theory and category theory — as more fundamental than arithmetic. That is, that you can explain arithmetic in terms of set theory, and set theory in terms of category theory. So one could argue that those are the more basic. But if we mean basic as in the first things anyone learns, yeah, it’s arithmetic. Definitely.

Kliban’s **Kliban Cartoons** for the 5th speaks of proofs. A good bit of mathematics is existence proofs, which is to say, showing that a thing with desired properties does exist. Sometimes they actually show you the thing. Such a “constructive proof” — showing how you make an example of the thing — pretty well proves the thing exists. But sometimes the best you can do is show that there is an answer. In any case, an example of a fish would convince all but the most hardcore skeptics that fish do exist.

I do at least one, and often several, Reading the Comics posts each week. They’re at this link. Essays that mention **Candorville** are at this link. Essays where I discuss **Dark Side of the Horse** are at this link. Appearances of **Frank and Ernest** should be at this link. Other essays with **Kliban** cartoons should be at this link.