Introduction: A Morning of Equations

At 4:55 the room starts to come up.

Not all at once. The lights climb by degrees, the way a theater comes out of blackout, and somewhere in the same instant The Verve starts playing. Bittersweet Symphony, every weekday, in a dark bedroom in Cuyahoga Falls. The song runs about six minutes. When it ends, SiriusXM's Chill takes over for fifteen more, and the room keeps getting lighter by an amount too small to catch in the act.

Then the whole second floor fills up with the voice of an old friend. Jeff Thomas, doing the Warmup Show on 92.3 The Fan, talking about the Browns to a man who is not yet upright.

That is the moment. Not the song and not the lights. The voice is the part that means something has to happen now.

So I lie there and I weigh it. Five more minutes, ten, or up. What is on the scale in those seconds is not what it looks like from outside. It is not laziness and it is not willpower. It is arithmetic, run fast and badly, by someone with his eyes closed.

How tired am I, which is a different question from how tired I feel right now, in the worst possible minute to ask it. How long does the shower take. How much air is built into the drive, not the ordinary drive but the bad one. Is there anything on the calendar this morning that punishes being late, or is this one of the many days where nobody would notice. And the hard one: are ten more minutes of this actually rest, or are they the same tiredness moved ten minutes down the road and handed to a man who now has to hurry.

Five or six terms, weighed and resolved and acted on before I have said a word out loud.

What I did not see for years was sitting in front of me the entire time. Everything up to that moment is a program. I wrote it. The song, the lights, the fifteen minutes of Chill, the handoff to the Warmup Show, all of it is an Alexa routine I built on purpose, running the same way every morning whether I am ready for it or not. The wake-up has been fully automated. It has stages. It escalates. It is a formula that executes while I am unconscious.

And the one thing I left out of it is the decision.

I automated the entire approach to the moment and then kept the moment for myself. There is no version of that routine that makes me get up. There never was. I did not build it that way by accident, though it took writing this book to understand why.

When I do get up, I get up. Across to the dresser, where the multivitamin and the prescriptions live because that is where they live. Good morning to Casey, who is a Yorkie, who is still asleep, and who does not acknowledge this in any way. Then the bathroom, and the shower goes on.

None of that was decided this morning. That sequence got solved once, a long time ago, and it has been running unexamined ever since. If somebody moved the dresser I would feel it immediately, which is the only time most of us ever find out we had a solution in the first place.

Downstairs it splits open again. Coffee some mornings. Other mornings a water bottle filled with Diet Pepsi, which I am aware is not a normal thing to admit, and which is a live question at that hour rather than a habit. Breakfast is a three-way: a granola bar, a Premier Protein, or nothing at all. I could not tell you now what I will pick tomorrow.

One of those sequences is compiled and the other is not. Same man, same kitchen, same fourteen minutes, and half of it runs without me while the other half gets computed fresh every single day.

It is still dark. Casey is awake now and has come to the top of the stairs to wait, because she knows that before I leave she goes down and gets fed. That is her formula, it has a dependency in it, and the dependency is me.

Then Route 8, and the Turnpike. Forty-three minutes on a good day, up to fifty-five on a bad one.

Most people carry the first number and call it the drive time. The second one is what actually decides anything, because what I am choosing when I walk out the door is how much of that twelve-minute spread I am willing to absorb and how much I am buying insurance against.

In the car it is Ken Carman and Anthony Lima, streaming, two more old friends talking about the same team Jeff Thomas was talking about ninety minutes earlier. Some mornings a podcast wins instead. I start between seven and seven-thirty, and which one it is comes down to whether there is an early meeting and how early I feel like being, My arrival time gets decided fresh, at speed, in the dark, standing at a dresser.

Twenty-five minutes early is twenty-five minutes spent. It bought something. Most people have never once asked what.

From 4:55 to the parking lot, call it a dozen questions, most with several terms, several turning on a forecast, one or two turning on a threshold I have never written down and could not state if you stopped me on the stairs and asked for it. All of them resolved somewhere between unconscious and merely distracted. Nearly all of them right. The day happened and nothing broke.

We often do not hear what was said. We hear what we were expecting to hear. The same thing happens to what we watch ourselves do. The expectation arrives first and the experience gets fitted to it afterward. You expect a morning, so you get a morning. All of it is running on schedule, right there in the open, sliding underneath the expectation without registering at all, the way a road you have driven ten thousand times goes under the car.

None of it felt like deciding. It felt like Tuesday.

There is a McDonald's on Howe Road, right off Route 8 south, one exit past my own.

It is not on the way. Reaching it means passing my own exit, coming off one exit later than I need to, and doubling back. Getting there costs me something before I have ordered anything.

The question is McDonald's or food at the house, and after enough repetitions I can tell you exactly what is in it. Let A be the perceived nutritional cost of the McDonald's against what I would eat at home. Let B be the perceived value of the energy and the time I keep by not cooking. Let C be the perceived cost of what it takes out of the week's eat-out budget. If A plus C is less than B, I get the double cheeseburger and fries.

That is not a metaphor for a decision. It is the decision, and it has an operator in it.

Each of those terms has to get a number from somewhere, and that is where the whole thing falls apart.

C comes closest to arithmetic. The budget is a real figure and this meal's share of it is a real figure. But the cost of spending it is not the dollar amount, it is whatever that money was otherwise going to do, which depends on what is still coming this week and what already happened last weekend. Same dollars on a Monday and on a Friday, two different costs. The number is real. Its weight moves.

A is coarser. I am not counting anything. I could not give you a calorie figure for either side, and looking one up would not change what I do, because that term is not doing quantitative work at all. It carries a rough sense of how the week has gone, and that sense moves the term. The same double cheeseburger costs more on the Thursday of a bad week than on the Monday of a good one, and nothing about the sandwich changed.

B is the loosest and usually the one that decides it. Half of it is measurable, since cooking plus cleanup is forty minutes or an hour and a clock can settle that. The other half is not measurable in any sense. The energy I keep by not cooking is a quantity I am certain exists, that I can rank confidently against yesterday, and that I could not put a unit on if you paid me for it.

Serious people have built machinery for exactly this. Ralph Keeney and Howard Raiffa published Decisions with Multiple Objectives: Preferences and Value Tradeoffs in 1976, and the idea underneath it is that when the things you care about come in different units you make them comparable by assigning weights, and once the weights exist the comparison is arithmetic again. Dollars in one place, minutes in another, and in the third slot something with no unit at all. That is precisely what my inequality does, and precisely where it hides its work, because the weights are the part I never stated.

Daniel Kahneman and Amos Tversky added something in 1979 that goes straight to the shape of it. In "Prospect Theory: An Analysis of Decision under Risk," they showed that people do not evaluate outcomes in absolute terms. We evaluate them as gains and losses against a reference point, and losses weigh more than gains of the same size. My equation carries two terms labeled cost and one labeled value. It was never a balanced scale. There is a thumb on it before the arithmetic starts.

Then there is the geography, which I ignored for years. I have never once run this equation in my kitchen. It does not occur to me at home. It occurs on Route 8 southbound with my own exit coming up, which means this was never a neutral comparison between two options. It is a challenge to a default, fired by where I happen to be, and the default wins every time it goes unchallenged. That asymmetry is nowhere in the formula. It lives in when the formula runs.

Which brings me to the word I wrote three times without hearing myself say it. Perceived nutritional cost. Perceived value. Perceived cost. The operation is arithmetic. The inputs are not.

So where does a perceived value come from? What sets the number? And if whatever sets it is not itself arithmetic, in what sense is any of this an equation?

I do not have an answer yet. It gets one eventually, and not in this chapter.

There was a keyboarding class at Sts. Joseph and John, which was my elementary school, and that is where the letter e stopped being a decision.

It was a decision then. Anyone who has watched a ten-year-old hunt for a key has seen the whole thing laid bare, slow and effortful and wrong a third of the time and consuming every bit of attention the kid has. I sat in that room and solved it. Then at some point I could not name within a year I stopped solving it, and my finger has been going there on its own ever since.

It never stopped being a decision. I just stopped being present for it.

That is the definition of automation this entire book runs on, and it applies to a person exactly as it applies to a system. A habit is a formula that was solved once and then compiled, so that it runs without being solved again.

Which answers a question I could not answer when I first went looking. The difference between a habit and a reflex is not speed and it is not effort, since both end up fast and effortless. The difference is whether anybody ever solved it. A reflex was never solved. Nobody taught you to blink when something comes at your eye. A habit was solved, on purpose, by you, and the solving is the part that goes invisible afterward.

If a habit is a compiled solution, then a habit can be opened back up and read and judged against the problem it was written for. A bad habit is not a character flaw. It is a bad answer that is still executing, and it is still executing precisely because nobody has looked at it since the day it was written. Everything I am going to say about systems is that same claim with more people in the room. The reason to get a rule out of somebody's head and into writing is not tidiness. It is that a rule nobody can read is a rule nobody can audit.

A man in Cambridge, Massachusetts, got here in 1890. William James, writing the chapter on habit in The Principles of Psychology, put the mechanism down flat: "habit diminishes the conscious attention with which our acts are performed." Then he drew the conclusion this whole book is downstream of. "The more of the details of our daily life we can hand over to the effortless custody of automatism, the more our higher powers of mind will be set free for their own proper work."

Twenty-one years later Alfred North Whitehead said it with his elbows out, on page 61 of An Introduction to Mathematics:

This is also the stretch of road where confident numbers get repeated that the research does not support.

You have heard that about forty percent of what you do in a day is habit. That traces to Wendy Wood, Jeffrey Quinn, and Deborah Kashy, "Habits in Everyday Life: Thought, Emotion, and Action," published in 2002, and what they actually found is narrower and more interesting than the round number. Two diary studies, participants logging what they were doing on an hourly signal. The first, seventy undergraduates over a single day, classified thirty-five percent of logged behaviors as habitual. The second, two hundred and nine undergraduates over two days, came in at forty-three. And habitual meant something specific: performed just about every day, and usually in the same location.

That is a real result and I am going to use it. It is also undergraduates, over one and two days, under a definition that is itself doing work, since same thing in the same place nearly every day will systematically miss the habits that travel with you and will scoop up routines that are merely scheduled. I will cite the studies. I am not going to cite the slogan.

The better evidence runs the other direction, toward building one on purpose. Peter Gollwitzer's implementation intentions are one of my own isms written in a psychologist's notation. Mine says almost everything in life can be reduced to a quantifiable formula, and that when X happens the user does Y. His says: decide in advance that if situation X arises, then I will do Y, and the behavior gets handed off to the situation instead of to whatever you have left in the tank at the moment. He laid it out in 1999, and with Paschal Sheeran pooled ninety-four independent tests finding an effect of medium-to-large size on goal attainment, d = .65.

It is a 2006 meta-analysis in social psychology, and the ten years that followed taught everyone in that field to read a number like that with the question of what never got published sitting right next to it.

Charles Duhigg's cue-routine-reward loop, from The Power of Habit in 2012, is the version most people have met, and it is a useful picture. It describes the shape a habit has, assembled mostly out of case narrative, and it is not a mechanism. Knowing the loop has three parts does not tell you why this loop formed and that one did not. I will take the shape and be straight about the rest.

My rule on what gets into an equation is this. Emotion is excluded as noise by default. It gets in as a weighted variable only when the result is going to significantly determine how I am going to feel.

That is an admissibility rule, and I mean it in the legal sense. The material is prejudicial more often than it is probative, so keep it out unless it bears on the question actually in front of you. Applied to most operational decisions I still think it is right. A rule like that exists because the alternative is a weight getting set by how somebody feels about a person rather than by anything the decision is about, and once that starts nobody can see it happening, including the person doing it.

Choose between two jobs on the numbers alone and you have solved the wrong equation, because the output of that decision is two years of how you feel at 4:55 in the morning. There, feeling is not interference with the data. It is the outcome variable. Leave it out and the model is not cleaner, it is wrong.

Now the case against me, at full strength, because you would find it on your own soon enough.

In 1994 Antonio Damasio published Descartes' Error, and in its third chapter, the one he called "A Modern Phineas Gage," he describes a patient he named Elliot. A tumor and the surgery that removed it damaged Elliot's ventromedial prefrontal cortex. His intelligence tested normal. Memory, language, attention, all intact. He could reason about a hypothetical as well as anyone in the room and generate options for it more or less forever.

That last part is what should be frightening. This is not a story about a man who got worse at thinking. By every measure of reasoning they could put in front of him, Elliot was fine. What he had lost was the feeling. He described the wreckage of his own life, the career and the marriages and the money all of it gone, with the flat detachment of a man recounting a stranger's afternoon.

And he could not decide anything anymore. Asked to pick one of two dates for his next appointment, he worked through reasons for and against each one at exhausting length and could not land on either.

Damasio's conclusion is that emotion is not interference in the decision process. In a wide class of decisions it is the process, the thing that assigns value to the options so that one of them can be preferred over the other. Take it out and you do not get a cleaner calculation. You get a man reasoning correctly, forever, about a choice of appointment dates.

Set that against Howe Road. Every term in my inequality carried the word perceived. Perceived by what? If Damasio has it right, the weights on A and B and C are not contaminated by feeling, they are set by feeling, and feeling is the only instrument I own that can set them. Which would mean my rule does not exclude emotion from the equation at all. It excludes emotion from the half of the equation I can see and leaves it in sole possession of the half I cannot, and that is arguably worse than letting it in the front door, because an input officially ruled out of scope is an input nobody ever inspects.

That one stays open through the whole book. Chapter 16 takes it up.

Not everything should get an equation.

My weekends have a formula and I run it on purpose. The grass needs cutting and the yard needs edging, and that needs daylight and a dry day, so it goes where the daylight and the dry day are. Groceries need the store open, so that has a window. Laundry runs in the background while something else happens, so it starts early and nothing waits on it. Floors get vacuumed and washed, the bathroom gets cleaned, the dishwasher gets run and emptied. Real dependencies, real windows, sequenced accordingly.

Then there is whatever is left when that is done, and I leave it completely unstructured.

Not because I ran out of energy for planning. Because scheduling it would not produce a markedly better weekend, and because the unstructured quality is the entire thing I was working toward. Putting that time on a grid would consume the good I was scheduling it for.

Two different claims get confused right here. One is that something cannot be formalized, that it is too subtle or too human or too whatever. I am generally suspicious of that one, and most times I have heard it, it turned out to mean nobody had tried hard enough. The other is that something should not be formalized, which is a completely different argument and a much better one. The leftover Saturday is the second kind. I could build the schedule. It would work. The result would be worse.

There is also a cost nobody prices. Every formula you build is a thing you now own. It has to be maintained, revisited when conditions move, and taught to anybody else who has to run it. A formula whose upkeep costs more than the benefit it throws off is a net loss even when it is completely correct, and being correct is exactly what makes that kind of loss invisible. You look at the rule, you confirm the rule is right, and you never ask whether it was worth having.

So the test is not whether something can be reduced to a formula. Nearly anything can. The test is whether the data-based version produces a genuinely better result than the instinctive one, net of what it costs to keep. Where it does not, leave it manual.

Which is what I built into an Alexa routine at 4:55 in the morning without understanding that I had. I automated the lights and the song and the Chill and the handoff to Jeff Thomas. I did not automate getting up. I could have. There are people who have. I left that one manual, every single morning, for roughly the reason I leave Saturday afternoon alone.

The obvious problem is that genuinely better is carrying enormous weight in that sentence, and the man who has automated nothing, who calls his avoidance intuition and his drift freedom, will say those exact words and mean something else entirely. A boundary that cannot tell judgment from avoidance is a permission slip. Chapter 15 is where that gets defined tightly enough to use, and turned into something you can run an actual decision through.

Three things send me back to an equation I have been running.

A bad result, which is the obvious one and the only one most people ever answer. New information, either from watching something behave differently than the equation predicted or from somebody making an argument I had not accounted for. And curiosity, with no bad result and nothing new, just the question of whether this still holds and whether it is as efficient as I assumed on the day I built it.

The third one gets skipped almost universally, and it is the one that finds anything, because of an asymmetry in the other two. A bad result forces an audit. A good result prevents one. Over any long enough stretch the only formulas that ever get examined are the ones that failed, and a formula quietly producing acceptable results is effectively immune from review. It goes on costing more than it needs to, or goes on answering a question that stopped being the right question two years ago, and the fact that nothing is going wrong is the thing protecting it.

Which is also why the second trigger has to include other people. If the only new information reaching an equation is what I notice myself, the audit is bounded by what I already look at. Somebody arguing with me is a search of a space I was not searching.

Revision, in practice, is almost never a rebuild. It is one term. A threshold moves. A variable that had been quietly carrying two things gets split into two variables. A term that was doing no work comes out.

And a revision has to leave a mark somewhere or it cannot be told apart from drift after the fact. That is where a personal habit and an organizational rule come apart. I can keep my own revisions in my head and mostly get away with it. A rule that ten people run cannot be revised in anybody's head, because the change has to reach ten people, and because the next person who asks why the rule says what it says deserves a better answer than nobody remembers. Changing an equation on purpose is revision. An equation that changed because nobody was watching is drift. From the outside they look identical, the thing simply not being done the way it used to be done, and they are opposites, because one has an author and a reason behind it and the other has neither.

All of that was one man's morning, which is exactly why none of it had to be written down. One reader, and the reader is also the author, and the variables can stay as private as they like.

Then somebody else has to run one of them, and everything changes.

Casey knows she gets fed before I leave. That formula has one dependency in it and the dependency is me, and it works perfectly because there is only one of me and I am never confused about what it means. Now hand a rule like that to another person. Feed the dog before you go. Before you go where, and how much, and what counts as before. Every one of those was a value I had been holding privately and never once had to say out loud.

That is not a communication failure. A communication failure gets fixed by saying it again more clearly. A definition failure survives being repeated, because the second telling contains the same undefined term as the first.

Multiply it out. Two people and one loose variable produce a disagreement somebody eventually notices. Thirty people and forty loose variables produce something else entirely, results that vary for no traceable reason, everybody involved following the rules, and the rules are not the problem. That is why organizations write things down, and whatever it feels like from the inside, it is not bureaucracy. It is the private threshold at a scale where nobody can hold all of them in their head anymore.

So the creed starts where it starts. Not with automation, which is the part everyone wants to talk about, but with definition, which is not.

My claim is that an organization is a morning with more people in it, the same system at a different scale, where the one genuinely new requirement is that the variables have to be said out loud because more than one person is running them. If that is right, everything after this scales cleanly and the creed is a single method operating at two magnitudes. It may not be right. A system does not just get bigger as you add people and parts. It changes in kind, and rules that work beautifully at one scale turn dangerous at another. Chapter 17 gives that argument its own chapter. My honest read is that the scale claim survives in a narrowed form. I am not certain, and I would rather say so here than protect it by never testing it.

The method takes the rest of the book, and it has a boundary, and the boundary has a boundary of its own. First, though:

Seeing them is all of it at this stage. Not optimizing them, not automating them, not being right about them. Just noticing that the ten minutes were a calculation. That the spread between forty-three and fifty-five is the actual decision and the forty-three is not. That the letter e was solved in a classroom at Sts. Joseph and John and has been running unsupervised ever since. That passing your own exit to reach a drive-thru is a fact about when the question gets asked and not about the sandwich.

And that the reason you and the person across the table cannot agree may have nothing at all to do with the argument, and everything to do with one word neither of you has ever defined. A startling share of the arguments I have watched were not disagreements. They were two faithful executions of the same rule, running on different numbers, by two people who each assumed the other meant what they meant.

The creed starts with definition, because nothing else runs until that is settled. And the people who spent their lives studying this are waiting in every chapter that follows. Including the ones who would tell me I have it exactly backwards.