Why a little pressure sharpens your work, and a little more wrecks it — and why the tipping point depends entirely on how hard the task is.
Performance rises as pressure, stress, or excitement (arousal) rises — but only to a point. Push past that point and performance falls, often sharply. The peak isn't fixed: on easy or well-practiced tasks you can handle a lot of pressure before it hurts you; on hard or unfamiliar tasks, the peak arrives early, and a seemingly moderate amount of stress is already too much. Most pressure-related failures aren't a motivation problem. They're a mismatch between how much arousal a task can actually absorb and how much it's being handed.
A quick gut-check before you reach for this model.
In 1908, psychologists Robert Yerkes and John Dodson ran mice through a simple maze, giving them small electric shocks of varying strength to motivate faster, more accurate runs. Mild shocks improved performance. Strong shocks made it worse. The relationship wasn't a straight line — it was a curve that rose, peaked, and fell. Decades later, that curve is still the reference point for how arousal and performance relate in humans.
Plot arousal (calm → wired) on one axis and performance (poor → excellent) on the other, and you get a bell-shaped curve. Too little arousal: you're under-aroused, under-motivated, sluggish, or bored — performance is weak. Moderate arousal: you're alert, focused, engaged — performance peaks. Too much arousal: attention narrows, working memory degrades, fine motor control and judgment suffer — performance collapses, sometimes fast.
This is the half of the law that actually makes it useful. For simple, well-rehearsed, or physical tasks, the peak sits far to the right — you can pile on quite a lot of adrenaline before it hurts, and extra arousal can even help (think of a sprinter or a weightlifter under a roaring crowd). For complex, unfamiliar, or judgment-heavy tasks, the peak sits far to the left — a surprisingly small amount of pressure is already past the summit, because the task needs spare working-memory capacity and calm attention that high arousal consumes first.
Try the simulator below — the same stress level is a helpful nudge for one kind of task and a performance killer for another.
A widely cited 2007 review (Lupien et al.) found that stress-hormone levels do track an inverted-U relationship with memory performance, lending real biological weight to the original behavioral finding. But researchers are candid that the exact mechanism isn't fully pinned down, and competing frameworks — flow psychology, reversal theory — argue the relationship between arousal and performance is more textured than one clean curve, especially for how arousal is *interpreted* (threat vs. challenge) rather than just how intense it is. Treat the inverted-U as a strong rule of thumb for calibrating pressure, not a precise formula.
Pick a task type, then drag the arousal slider and watch where performance actually peaks.
| Setting | Under-aroused | Sweet spot | Over-aroused |
|---|---|---|---|
| Sales pitch / live demo | Flat delivery, no energy, loses the room | Alert, crisp, reads the room, handles questions well | Talks too fast, forgets key points, over-explains out of nerves |
| Surgery / high-stakes procedure | Complacent, misses small cues, slow to react | Focused, steady hands, calm scanning for problems | Tunnel vision, hands shake, decision quality drops under time pressure |
| Competitive sport | Sluggish, un-motivated, "flat" game | "In the zone" — fast, loose, confident | Choking — overthinking routine plays, muscle tension, panic |
| Exam / certification test | Careless errors from boredom or low stakes | Sharp recall, good pacing, catches own mistakes | Blanking, re-reading the same question, running out the clock |
| Customer support under SLA pressure | Slow resolution, little urgency | Efficient triage, clear communication | Rushed troubleshooting, missed root cause, reopened tickets |
| Software incident response | Low urgency, slow to escalate a real outage | Clear-headed triage, calm comms to stakeholders | Panicked rollback decisions, tunnel-visioned on one hypothesis |
Don't reach for this model when the problem is chronic rather than momentary — ongoing burnout, sustained overwork, or a mental-health condition need a different kind of response entirely, not a tighter arousal dial. Don't use it to retrofit a justification for a high-pressure culture after the fact. And don't use it as your only lens on a performance problem — skill gaps, unclear instructions, bad incentives, or plain exhaustion can all look like "wrong arousal level" without actually being one.
A team kept blowing investor demos despite flawless rehearsals. Here's how calibrating arousal — instead of just "telling people to stay calm" — actually fixed it.
Result: the scripted segment stayed energetic, the Q&A got noticeably calmer and sharper, and the team stopped treating "nerves" as a single, undifferentiated problem to be willed away.
Describe a real situation. This builds a ready-to-use prompt you can hand to any AI assistant (or think through yourself) to diagnose where you're sitting on the curve.
A portable reference you can hand to any AI assistant, or keep for yourself, so this model is one copy-paste away whenever you need it.
Describe what you're dealing with, and get a tailored prompt that applies the Yerkes-Dodson Law directly to it.