A Field Guide
On plans, systems, and the places reality gets creative
"The universe tends to unfold as it should."
— Kumar Patel, Harold & Kumar Go to White Castle, 2004
The Premise
It is not the belief that things will work out as planned. It is the quieter conviction that something is happening in the parts of the situation you cannot currently see.
Regular optimism looks directly at the problem and says it will be fine. Peripheral optimism doesn't look at the problem at all. It trusts that the system is doing more than the visible parts suggest.
It doesn't require you to feel good about the failure. The rats were still a failed mission. Peripheral optimism just holds open the possibility that the story isn't finished — and the interesting part may be arriving from the side.
"Learning from failure" requires wisdom and attention. Peripheral optimism doesn't. It proposes that failure does things whether you're paying attention or not. Reality is quietly composting your failures without asking permission.
When you've warned someone and they didn't listen and you're watching it unfold — this is the off-ramp. You don't need to manage consequences. The failure is going somewhere without your supervision. Step back. The periphery is handling it.
"Nothing is fully wasted.
It just invoices differently than expected."
Classification System
Naming the type of failure changes what you do next. Misidentify it and you apply the wrong response — fire people instead of redesigning the system, treat noise as signal, or shut down something quietly working.
The incentive produces the opposite behavior. The system is rational. The design was wrong. People optimize precisely for what you measure, which is almost never what you actually want.
Lesson: People optimize for what you measure, not what you want.
One intervention destabilizes a system balanced in ways nobody mapped. The sparrows ate the locusts. Nobody thought to ask what the sparrows were doing before removing them.
Lesson: Everything is connected to something you haven't met yet.
A brilliant solution to the last problem. The line held perfectly. Germany went around it, as in the previous war. Preparation had hardened into assumption.
Lesson: Preparation hardens into assumption.
Observation changes the thing being observed. The researchers changed the lighting and productivity rose. They changed it back. Still rose. The measurement was the intervention.
Lesson: You are always part of the experiment.
Failure that succeeds through an unintended mechanism — often better than success would have. The rats didn't explode anything. They just had to exist in Nazi imagination.
Lesson: Energy doesn't disappear. It invoices differently.
Correct answer, rejected. He proved handwashing saved lives in 1847. The establishment refused him. He died of the same infection. The practice became standard decades later.
Lesson: Being right isn't sufficient. Timing and framing are load-bearing.
Nobody told the king. Everyone knew the ship was unstable. It sank 1,300 meters into its maiden voyage. Still within sight of the crowd. Hierarchy suppressed information until the sea resolved it.
Lesson: The most dangerous person in the room is the one nobody will correct.
Solution that worked in the short term and catastrophically in the long term. The US government paid farmers to plant it. It now smothers millions of acres. The erosion problem is fine.
Lesson: Short-term metrics are a loan against long-term reality.
The correct and working solution, dismissed because it looked ridiculous. The pigeons guided missiles accurately. The military couldn't say the words. They were right to laugh and wrong about everything else.
Lesson: Credibility and validity are not the same thing.
Historical & Business Archive
Cases from history, ecology, and the modern workplace. Filter by context or type. Each one is a system responding to a plan in ways the plan didn't anticipate.
The SOE stuffed dead rats with plastic explosives, planning to place them near furnaces in German factories. The shipment was intercepted before any were deployed. But German intelligence was so alarmed they ordered sweeps of all dead rats across occupied Europe. Soldiers were instructed to treat every dead rat as a potential explosive. The mission failed. The paranoia cost far more time and resources than the bombs ever would have.
The colonial government offered a bounty for every dead cobra. Enterprising locals began breeding cobras for the reward. When the program was cancelled, breeders released their now-worthless stock. The cobra population ended up higher than before the intervention. The incentive had done its work with perfect rationality toward the wrong target.
Mao mobilized an entire country to kill sparrows for eating grain. Hundreds of millions were killed. The following year, locust populations exploded — previously controlled by the sparrows. The famine contributed to tens of millions of deaths. Sparrows were quietly removed from the pest list and replaced with bedbugs.
Researchers changed lighting conditions. Productivity rose. They dimmed the lights. Productivity rose. They changed break schedules. Rose. They changed them back. Still rose. The workers knew they were being observed, and being observed was the intervention. The study set out to measure lighting and discovered something about attention itself.
Sweden's most powerful warship sank 1,300 meters into its maiden voyage, still within sight of the assembled crowd. Engineers knew it was unstable. Nobody told the king. Recovered three centuries later nearly intact, it now sits in a museum as one of Sweden's most visited attractions. The catastrophe became the treasure.
Ignaz Semmelweis proved handwashing between autopsies and deliveries saved lives. The medical establishment rejected him. He died in an asylum, likely of the same infection he'd spent his career preventing. Hand-washing became standard practice decades later. He was right. He was also early, which functioned as being wrong.
France built an impenetrable fortified line along the German border after WW1. Germany went through Belgium instead, exactly as in the previous war. The Maginot Line was never breached. It successfully defended its section. France fell in six weeks. The preparation was thorough, competent, and oriented entirely toward the previous crisis.
Barbara Streisand sued to remove an aerial photo of her home from a coastal survey. Before the lawsuit: downloaded six times, twice by her lawyers. After the lawsuit made news: downloaded 420,000 times in a month. The suppression mechanism became the most efficient distribution system available. The legal attempt is now named after her.
Management set aggressive cross-selling quotas — every customer should hold eight products. Employees couldn't hit them legitimately, so they opened approximately 3.5 million fake accounts in real customers' names without authorization. The metric was being hit perfectly. The underlying goal — actual customer relationships — was being actively destroyed. The quota measured the wrong thing so precisely it automated fraud.
Every team was required to rank employees and a fixed percentage had to be rated "poor" regardless of actual performance. If you were on a team of stars, someone was still getting fired. Employees stopped collaborating, stopped sharing knowledge, actively undermined colleagues. The review system designed to identify underperformers manufactured underperformance as a survival strategy. The company's innovation stalled for most of a decade.
McNamara couldn't measure who was winning the Vietnam War meaningfully, so he measured kills per engagement. The metric became the mission. Soldiers reported inflated counts. Officers demanded higher counts. The war looked like it was being won in every report, right up until it wasn't. He measured what could be counted and assumed what couldn't be counted didn't matter. It did.
Net Promoter Score surveys ask how likely customers are to recommend the company. Companies began coaching customers before sending surveys — call center workers told customers "if you can't give us a 10, please call us back first." The score became a measure of how well you coached survey responses, not how good the product was. The metric was measuring the metric-collection process.
Mark-to-market accounting let Enron book projected future profits as current earnings the moment a deal was signed. Technically legal. The earnings looked extraordinary. The actual cash was not there. The model measured optimistic futures as present reality until it couldn't anymore. The audit confirmed the reports. The audit was measuring the reports, not the company.
Gap replaced its iconic logo with a generic redesign. The backlash was immediate and enormous — social media mockery, widespread ridicule, design community outrage. Gap reverted within a week. The failed rebrand generated more awareness of the brand's identity, more affection for the original logo, and more press coverage than any successful redesign would have produced. The disaster was the campaign.
Companies spent enormously on computers expecting immediate productivity gains. Robert Solow observed in 1987: "You can see the computer age everywhere except in the productivity statistics." The investment was right. The payoff took twenty years to materialize as organizations slowly restructured themselves around the technology. Everyone measuring year-over-year looked like they'd wasted the money. They hadn't. They were just early, which functioned as being wrong.
Trains were measured on whether they arrived within ten minutes of schedule. Trains started being scheduled for longer journey times so they could arrive "early" and count as on time. The trains didn't get faster. The schedule got slower to accommodate actual performance. The metric was perfect. The service degraded. Passengers noticed what the metric didn't.
Composite Scenarios
No company names. No org charts. Just the pattern, running on schedule, in meetings and performance reviews near you right now.
The metric was calls resolved per hour. The team started closing tickets without solving them. Transferring unresolvable calls in circles. Hanging up on difficult customers. Average handle time: excellent. Customer trust: quietly collapsing. The dashboard looked healthy the entire time. A second metric was introduced to measure the first metric's failure. It will also be gamed.
What type: Cobra Effect, likely cascading into a second one.
What not to do: Add another metric. Redesign the incentive structure instead.
Everyone in the room knew the timeline was impossible. The estimates had been padded. The dependencies were unmapped. The project manager knew. The developers knew. The QA team definitely knew. The VP asked if it was on track and everyone said yes. The deadline was missed by three months. The post-mortem identified "communication issues."
What type: Vasa — hierarchy suppressing information until reality resolves it.
What not to do: Run another post-mortem. Build a channel where bad news travels fast.
The last reorg fixed the problem. Siloed teams, slow decisions, communication breakdowns — all addressed. The new structure was elegant. It solved exactly the problems the previous structure had created. Eighteen months later, the same symptoms appeared in different departments. Another reorg was scheduled. This one would also be elegant.
What type: Maginot — preparation hardened into assumption.
What not to do: Build a permanent structure for a dynamic problem.
Someone documented a systemic process failure and submitted it through proper channels. Nothing happened. They mentioned it to a colleague. That colleague told their manager in a different department. That manager had seen the same pattern. The documentation circulated informally for three months. By the time leadership addressed it, five departments had already changed their behavior. The official complaint failed. The fix spread anyway.
What type: Rat Bomb — failure that travels through unintended channels.
What to notice: The thing that didn't work officially may still be working.
Leadership introduced a real-time performance dashboard visible to the whole floor. Output increased immediately. The dashboard was expanded. Output increased again. Six months later, output returned to baseline. The increase hadn't been improvement — it had been performance during observation. The observation was the intervention. The dashboard was now wallpaper. A new dashboard was proposed.
What type: Hawthorne — measurement changes behavior temporarily, not structurally.
What not to do: Add more dashboards.
Someone proposed a different approach to the quarterly process. The data supported it. It made three senior people feel implicitly criticized for how they'd been doing it. The proposal was tabled. Two years later, an external consultant recommended the same approach. They were paid forty thousand dollars. It was implemented immediately. The original proposer is still there.
What type: Semmelweis — being right requires a receptive system.
What to remember: Being right isn't sufficient. Timing and framing are load-bearing.
Interactive Diagnostic
Describe a situation — workplace, organizational, or personal — and identify which pattern is operating. Knowing the type changes what you do, and what you stop doing, next.
Core Observations
Plans are linear. Systems have ten thousand variables responding simultaneously. The system always has the last word. This is not pessimism — it is the source of every unexpected silver lining in this archive.
Metrics, quotas, body counts, NPS scores — all maps. The territory continues doing what it does regardless of what the map reports. The moment a map becomes authoritative, it starts diverging from the ground it claims to represent.
Not malicious. Not random. Operating at a resolution higher than the plan was written at. The Nazis inspected dead rats. The Vasa became the museum. Reality is not cruel — it is thorough.
The sparrows and the locusts. The mongooses and the birds. The cobra bounty and the cobra farms. Every intervention touches something invisible until it isn't. The absent variable is almost always the most important one.
It goes somewhere and does something. Not always useful. But more often than expected, the energy from a failed plan arrives somewhere unintended and does work there — without anyone noticing, without anyone taking credit. The thing that didn't work is still working. You just can't see it yet.
The Personal Application
You saw it coming. You said something. They didn't listen. Now you're watching the cascade from the side — oscillating between vindicated and frustrated, carrying the weight of being right in a room where being right didn't matter.
Peripheral optimism is the off-ramp. Not because you were wrong to warn them. Because the failure is already going somewhere without your supervision. The system is responding. The consequences are accruing. The thing you saw is doing its work.
You don't need to manage it. Reality doesn't require your help enforcing consequences. It is extraordinarily good at that without you.
The rats didn't need anyone to explain what happened. The paranoia spread on its own schedule, through channels the SOE never accessed. The failure traveled. It arrived somewhere useful. Nobody was there to see it.
Trust the parts you can't see. They've been doing the work the whole time.