Beware the Mean

https://stephen.sh/posts/beware-the-mean

by sjwhitworth • 6 years ago

94 36 6 years ago

kqr4• 6 years ago

Other, related points:

- With heavy tails, the sample mean (i.e. the number you can see) is very likely to underestimate the population mean.

- With heavy enough tails, higher moments like variance (and therefore standard deviation) do not exist at all -- they're infinite.

- Critically: With heavy tails, the central limit theorem breaks down. Sums of heavy-tailed samples converge to a normal distribution so slowly it might not realistically ever happen with your finite data. Any computation you do that explicitly or implicitly relies on the CLT will give you junk results!

wodenokoto2• 6 years ago

Can you elaborate on the part about sample vs population mean?

The way I see it; in these scenarios you aren’t looking at the sample mean. There is no reason to sample your customer base to get an estimation of your average revenue. You can calculate from the entire population.

paulsutter• 6 years ago

Your current customer base is just a sample of your total market, for example. Next year’s (larger) customer base will be slightly closer to the whole set (so different) but still just a sample.

j7ake• 6 years ago

An example of your infinite variance case, sums of cauchy distributions don't even converge to a normal distribution.

contravariant2• 6 years ago

With heavy enough tails even the mean might not exist.

mumblemumble2• 6 years ago

Every non-empty set of numbers has a mean.

Perhaps you mean that, in a multimodal distribution, the mean might not resemble any individual member of the population?

mlyle2• 6 years ago

I would quibble some here. When we look at revenue, I agree: ignore the mean. If there's a whole bunch of people not paying you anything, that's OK... Look at the 50th and 90th percentile.

But profit, and similarly costs? Your mean customer better be profitable, or you won't be. How much the people on the left of the graph cost you is important.

Part of this is definitional, too. Do you include that far left part of the graph where people are not really paying you as a "customer"?

stdbrouw• 6 years ago

Part of this is definitional, too. Do you include that far left part of the graph where people are not really paying you as a "customer"?

Exactly. Often when the mean is problematic, it's because there's multiple subgroups in the data with very different behavior, and when you split those up, you can learn a lot more from the numbers.

andreilys• 6 years ago

A “mean” profitable customer would be misleading. You could see that 99% of your customers lose money for you, while 1% earn all the profit (aka fat tail)

The point is you should be careful about reducing huge swaths of data to one datum. It often hides the more interesting insights.

mlyle2• 6 years ago

A “mean” profitable customer would be misleading.

In no place did I suggest "mean profitable customer" as a metric. I sort of said the opposite.

At the same time, in a traditional industry we wouldn't consider people with an evaluation license or sample or who came in to wander and see our wares a "customer". They're a lead and have a cost associated with them.

The point is you should be careful about reducing huge swaths of data to one datum. It often hides the more interesting insights.

Sure. At the same time, if you dance with 30 numbers looking for insight, pretty soon you're practicing qualitative, wishy-washy innumeracy. The discussion is about KPIs, which are all about "reducing huge swaths of data to one datum" but also absolutely essential to run a real business day to day.

Monitor more than one, and keep your eyes open for where they go wrong, and be ready to change what you do.

fmajid• 6 years ago

The Iranian civilization can draw continuity to Susa, circa 3000BC, further than China. The Mesopotamian and Indian civilizations are older still but broke continuity.

RubenvanE• 6 years ago

I think you meant to reply to this post https://news.ycombinator.com/item?id=22166846. :)

ImaCake• 6 years ago

If the author is seeing this thread; I couldn't find an RSS feed for your site. I don't know if they are difficult to setup, but if it's very little effort, I would appreciate seeing what you post next :)

As for the waryness about the mean. A lot of people much further behind than thinking of different distributions. Even something you assume is normal distributed needs a mean and a variance! As for visualising, histograms are incredibly underrated tools. You can infer a lot of information by just looking at a distribution.

edf13• 6 years ago

You could always try to contact him direct:

https://stephen.sh/about

sjwhitworth• 6 years ago

Sorry, I do need to set this up. It's just a bunch of Markdown at the moment. Thanks for reading!

EliRivers• 6 years ago

The mean is misleading. The median is misleading. The mode is misleading. Any reduction of a range of data to a single representative datum is misleading.

However, the fight back against providing something a bit more meaningful than a single value can sometimes be quite strong.

I try hard to provide software estimates as probability distributions, but when someone sees a line with a probability peak somewhere around two days (could be really simple), and then a wide hump somewhere around two weeks (if it's not simple, it will mean a significant rewrite), with a very low line between them and then a long, long tail off to several months, it is not well-received.

I can see their point; they're trying to plan things, and the whole system is set up to work with single numbers. If everyone provided probability graphs for their estimates, and we had a tool that could then combine them and deliver the net probability graph of the combined pieces, I expect they'd be a lot more amenable.

MakeUsersWant• 6 years ago

What is your solution in practise?

pif• 6 years ago

they're trying to plan things

Each divorce starts with a marriage. Each delay starts with a plannig.

Izkata• 6 years ago

The mean is misleading. The median is misleading. The mode is misleading. Any reduction of a range of data to a single representative datum is misleading.

For anyone who hasn't seen it before, Anscombe's Quartet is a nice visual of this (and actually goes a bit further, showing reduction in general can be misleading, not just to a single point).

https://en.wikipedia.org/wiki/Anscombe%27s_quartet

petercooper• 6 years ago

I was just coming here to post that. A very neat (and surprising) thing! :)

teodorlu• 6 years ago

Nassim Taleb greatly expands on this point in Antifragile. For a freely available, techical argument, check out Doing Statistics Under Fat Tails[1].

[1] https://www.fooledbyrandomness.com/FatTails.html

theophrastus• 6 years ago

This is a worthy posting, particularly as so much becomes iterative statistics in "A.I." clothing. The two old (slightly hackneyed) counter-examples which are popular in lectures about measures of the central tendency are:

- One is trying to get a sense of the common sort of income in a room and then Bill Gates wanders in. Suddenly the average income becomes an amount which no one experiences.

- What is the average number of testicles in the human population? That computed central tendency is quite rare.

kqr4• 6 years ago

The second one doesn't seem that bad to me. The number can still be used to answer common questions like, "Assuming it takes 10 seconds to tickle a testicle, how long do I have to tickle testicles if I want to tickle every testicle in my apartment complex."

Sure, you have to know what the mean means, but it's still a useful number. Your first example is more indicative of the problem, IMO.

PaulHoule• 6 years ago

The mean is not so bad for many purposes because it is an expectation value.

If you add up your revenue, subtract your expenses, and divide by the number of customers that gives you a real profit number. (Condition how you define revenue & expenses) If that number is negative or positive it is meaningful.

The median on the other hand has a different set of problems. If you are running a game like Fate Grand Order you'd better cultivate the guy who spends $70k because he has to "catch them all". The median player probably pays little or nothing, but the guy who sells ero comics at Comiket complains about what it costs to get (say) Saber Bride, but it is worth more to him than it is to the medium.

Mean and median are terrible numbers to use for latency; what drives you nuts with your computer being unresponsive is not the median latency, but the 99% latency.

kqr4• 6 years ago

The mean is not so bad for many purposes because it is an expectation value.

Implicit assumptions:

1. The sample mean is an accurate estimator of the expectation.

2. The expectation is a useful number.

Both of these are false surprisingly often; an example is one you're mentioning: latencies.

pototo666• 6 years ago

I came across too many people who value mean soooooo much in the analysis. Well, some of them made mistake and the project died. Hypothesis: heavy reliance on mean increases the probability of failure in internet industry. This reminds of PG's essay mean people fail: http://www.paulgraham.com/mean.html

Pun intended :)

Beware the Mean

Related Stories