Critical Section

estimating in meatspace

Thursday,  08/04/05  11:24 PM


I am really good at estimating.  By which I mean, I can estimate anything, even if my estimates are not accurate.  If you ask me for an estimate – on anything – I can give you one.  I never let absence of facts stand in my way.

However, many people are horrible at estimating.  You ask them for an estimate – on anything – and they can’t do it.  They don’t know where to start.  Even if they should be able to estimate something, based on their experience and knowledge and the availability of facts, they just can’t do it.  It isn’t that they don’t want to commit – they might say that’s why, though – it is because they honestly can’t make an estimate.

In the past I have found this frustrating.  I think to myself, “if I had your experience and these facts, I could form an estimate immediately”.  Experience has taught me that getting mad and putting pressure on someone to give you an estimate when they think they can’t doesn’t help.  (This is especially true if the “someone” in question is your wife or daughter :)

But recently I’ve found a great trick.  You can help people form estimates by using binary searchingPeople are much better at comparisons than they are at estimating.  This is true even though all you need to do to form an estimate is iterative comparisons.

Binary searching

The technique of binary searching is well known in computing, it is part of every freshman computer class.  It is a routine technique; suggesting that a programmer use binary searching is like suggesting to a fish that they swim.  However non-nerds might not be familiar with the technique so here’s a quick example.

Pick a number between 1 and 8.  Got it?

Okay, I can guess your number with three questions.  Here we go:

#1) Is your number bigger than 4?

Yes - #2) Is your number bigger than 6?

Yes – #3) Is your number bigger than 7?

Yes – your number is 8!

No – your number is 7!

No – #3) Is your number bigger than 5?

Yes – your number is 6!

No – your number is 5!

No – #2) Is your number bigger than 2?

Yes – #3) Is your number bigger than 3?

Yes – your number is 4!

No – your number is 3!

No – #3) Is your number bigger than 1?

Yes – your number is 2!

No – your number is 1!

See how it works?  Each question is a comparison, and each result divides the number of possible answers in half.

Estimating by binary searching

Okay, so let’s say you want someone to give you an estimate.  Instead of asking them directly for the estimate, ask them to do a series of comparisons, each of which cuts the possible answers in half.  Pretty soon you have a reasonable estimate.  Here’s a real example:

You’re talking to a pool repairman and you want to know how much a particular repair will cost.  He says he has no idea.  But you know he should have a good idea, based on having been a pool guy for years.  Here’s the conversation:

You:  Do you think this will cost more than $100,000? 
(NOTE: start with an obvious upper limit.)

Him:  What!  Of course not.

You:  Whew!  Okay, well, will it cost more than $50,000?

Him:  No way.  Not nearly that much.

You:  Okay, well, will it cost more than $25,000?

Him:  No, this isn’t that bad.  I didn’t mean to scare you, it will only take a few days to fix.

You:  Okay, well, will it cost more than $10,000? 
(NOTE: pick round numbers!!!)

Him:  Hmmm….  No, it won’t cost that much.

You:  Will it cost more than $5,000?

Him:  Let me think.  No, I don’t think so.  I did a job a couple of weeks ago which was just like this, and it was less.
(NOTE: Resist the urge to say “why didn’t you just say that in the first place”.  Don’t snatch defeat from the jaws of victory :)

You:  Okay, so will it be more than $2,000?

Him:  Oh, yeah, I have to get a new framitz, and my guys have to decombobulate the qwork, it will be more than that.

You:  Yeah, I was afraid of that.  Okay, well will it be more than $3,000?

Him:  It might be.  I’ll try to stay under that, but with these things you never really know how much they’ll cost.

See how it works?  This guy really knew the right estimate was “about $3,000”, but he didn’t know he knew, and he didn’t know how to get there.  By binary searching with comparisons you helped him unearth his knowledge.

Be careful always to phrase each question as a comparison.  Don’t ask “will this be about X?”, instead ask, “will this be more than X?”  I don’t know why, but humans do better with “more than” -type questions.

This technique even works within yourself.  Let’s say you have a project to do.  (You need to decombobulate a qwork.)  You can’t form a decent estimate.  Ask yourself – will it take more than a month?  More than two weeks?  More than a week?  More than two days?  More than eight hours?  Etc.  You know more than you know!

I hope you find this as useful as I have!

(Did you find it more useful than...  just kidding :)

this date in:
About Me

Greatest Hits
Correlation vs. Causality
The Tyranny of Email
Unnatural Selection
On Blame
Try, or Try Not
Books and Wine
Emergent Properties
God and Beauty
Moving Mount Fuji
The Nest
Rock 'n Roll
IQ and Populations
Are You a Bright?
Adding Value
The Joy of Craftsmanship
The Emperor's New Code
Toy Story
The Return of the King
Religion vs IQ
In the Wet
solving bongard problems
visiting Titan
unintelligent design
the nuclear option
estimating in meatspace
second gear
On the Persistence of Bad Design...
Texas chili cookoff
almost famous design and stochastic debugging
may I take your order?
universal healthcare
triple double
New Yorker covers
Death Rider! (da da dum)
how did I get here (Mt.Whitney)?
the Law of Significance
Holiday Inn
Daniel Jacoby's photographs
the first bird
Gödel Escher Bach: Birthday Cantatatata
Father's Day (in pictures)
your cat for my car
Jobsnotes of note
world population map
no joy in Baker
vote smart
exact nonsense
introducing eyesFinder
to space
where are the desktop apps?
still the first bird
electoral fail
progress ratches
2020 explained