What Failure Knows That Success Cannot Teach

On failure: we quit every day and start up again the next day.

The name that is not there

The lineup went up on the wall of the old Ky Ebright boathouse, and some mornings the lineup did not go my way. Nobody said anything. The names were just there, or not.

Imagine the seconds after you find the wall. You read the names top to bottom, then again, slower, as if a second pass could change them. Nobody looks at you, and that kindness is hard to take. You pick up a bag, anything with weight, so your hands have a job. The meaning comes later, and it comes dressed as a verdict.

The wall was not the first hard thing. At ten I had a stroke, and my left side went weak. That was not a failure. It was something that happened to a ten-year-old body. Any sense of failure came later, from the meaning given to it, and that is a separate thing from the event.

Everything I did in sport afterward I did with a left side that was weaker than my right. I went on to row competitively anyway. Rowing is a strange choice for a body like that: every stroke has to match on both sides of the boat. Years later I was up at 5:00 and at the boathouse by 6:00 like everyone else.

I am not telling you that as a triumph. I tell it because it taught me early that a limit can be read, not only felt. A weaker side says where the work is. It does not say who you are.

Why does the same lost seat race make one rower faster by June and send another home for good? I do not think the difference is toughness. I think it is what each of them does in the hour after reading the wall: react to the failure, or interpret it.

A coach who could not win the big one

The best-known crew in rowing history was built by a man who spent years losing. As Daniel James Brown tells it in The Boys in the Boat, Al Ulbrickson took over the University of Washington crew in 1928 at twenty-four, and by the summer of 1935 the national varsity title had still not come.

That June at Poughkeepsie, on the Hudson, his varsity led into the last mile and then was passed, first by California and then by Cornell. Third.

The next season he rebuilt, reshuffling lineups until one combination ran. That boat never lost a race together, through Berlin. The losses of 1935 did not make that boat. What Ulbrickson took from them may have. He read 1935 as information about lineups, not as a verdict on himself.

What failure does to learning

Start with the bad news. In 2019 Lauren Eskreis-Winkler and Ayelet Fishbach ran five experiments with more than 1,600 people. Each question had only two answers, so being told "wrong" carried exactly as much information as being told "right." People still learned less from failure. In one study telemarketers answered a customer-service quiz; on the retest, those who had been told when they were right scored 62 percent, those told when they were wrong scored 48. The researchers' reading is ego threat: failure feels bad, so we tune out. When people watched someone else fail, they learned from it just as well as from success. These are quick quiz tasks, so I treat the size as a hint.

The sting is calibrated, and it is not to the result. In 1995 Victoria Medvec, Scott Madey and Thomas Gilovich had raters score footage of medalists from the 1992 Barcelona Games on a ten-point scale from agony to ecstasy. It was a small sample, about three dozen athletes. Bronze medalists looked happier than silver medalists, both right after the event and on the podium. The usual account is counterfactual thinking: silver may compare up, to gold, and bronze may compare down, to no medal. In 2021 William Hedgcock and colleagues repeated the test with software on medal-stand photos of 413 athletes from five Games and found the same pattern, sharpest in silver medalists expected to do better. That is an average pattern, and the counterfactual story is an explanation, not a measurement of any one athlete. I hold that the degree of our emotions is in direct relationship to how much we put in and how much we care. The size of the feeling is a reading of the stake.

The "if only" can be fuel or a loop. Neal Roese's work on counterfactual thinking found that imagining how things could have gone better, an upward counterfactual, can raise intentions and performance. In one of his studies, people who did this after a block of anagrams did better on the next, largely by finding a strategy. The work is mostly lab tasks, and the benefit fades when there is no next chance. Repeated without a plan, the same thinking is linked to rumination and low mood. "If only" helps when it ends in "next time."

Shame and guilt split the same failure in two. June Tangney's review of the moral emotions, building on Helen Block Lewis, draws the line between shame, which is about the self, and guilt, which is about a specific behavior. Guilt tends to move people toward repair. Shame tends to move them toward hiding, defensiveness and sometimes anger. "I raced badly" can be worked on. "I am a bad rower" has nowhere to go.

That is the line this whole piece turns on. Failure as information says: this happened, and here is what it tells me about the next attempt. Failure as identity says: this happened, so this is what I am. The event can be identical. Only the first one leaves you anything to do.

Failure contains information. It does not guarantee understanding. What we learn depends partly on whether we can look at what happened without turning the result into a verdict on who we are.

Kindness toward yourself is not letting yourself off. Amber Mosewich and colleagues randomized 51 self-critical women varsity athletes to a seven-day self-compassion program or an attention control. The program paired one teaching session with writing exercises. Self-criticism, rumination and concern over mistakes dropped, and the gains held at four weeks. It is one small study, in one population, with self-report outcomes, so I treat it as promising.

Third place, second place

Read each row left to right: the athlete who finished higher looked less happy at both moments, because the feeling was measured against the gold that almost happened, not against the medal in hand.

What each failure is reporting

Each of these feelings arrives with a message and an urge. The message is usually worth keeping. The urge is the reaction. This is my working reading, drawn from the research above, not a diagnostic key.

The last column is the whole skill in miniature. Keep the signal, decline the instruction.

Take regret. Useful regret ends in a plan: "I rushed the slide when I got tired, so next piece I count the first ten." Rumination replays the same moment and ends in a verdict: "I always do this." The first has a next step. The second only has a loop.

Get up again

There is a cheerful version of this article that I do not believe. It says failure is a gift and every setback makes you stronger. Failure is not a gift by default. It is raw material, and raw material can be wasted.

Errors can teach, under the right conditions. Nina Keith and Michael Frese pooled 24 studies of error management training, where trainees are told to expect mistakes and use them. It beat other training by a moderate amount overall, and by a large amount on new problems unlike the ones practiced. Nearly all of those studies taught software, so I hold it as promising. Robert and Elizabeth Bjork's desirable difficulties point the same way: some kinds of practice that feel harder, such as spacing and mixing, often build more durable learning, even while it looks worse on the day.

Other claims have shrunk. Growth-mindset interventions showed a small average effect across 29 studies, d = 0.08, and a 2023 meta-analysis by Macnamara and Burgoyne found the best-designed studies showed almost nothing. Post-traumatic growth is real for some people, but Eranda Jayawickreme, Laura Blackie and colleagues have pointed out that most of the evidence rests on people's own looking back, and a later study of over 800 adults found that feeling you had grown did not predict measured change.

Put the findings side by side and the conditions for learning start to show. The feedback has to be specific enough to act on. There has to be a next chance, or the "if only" turns into a loop. The person has to be safe enough to look, because ego threat makes us tune out. And distance helps: we learn more easily from someone else's failure than from our own.

The Desert Fathers knew this, in their way. A brother came to Abba Sisoes and said he had fallen. What should he do? The old man told him: "Get up again." He had, said the brother, and fallen again. Get up again and again, said Sisoes, and he did not set a limit.

Epictetus adds the piece that turns getting up into learning. In George Long's translation: "Men are disturbed not by the things which happen, but by the opinions about the things." He goes on to say the uninstructed person blames others for their trouble, the beginner blames himself, and the one whose instruction is complete blames neither. I read that as the shame-to-guilt move with no blame left at all. Not "they cost me the seat." Not "I am not good enough for the seat." Just: here is what happened, here is what it says.

These traditions and the quiz studies do not prove each other. They rhyme. Both say the self gets in the way of what the failure is trying to tell you.

When I think of my left side at ten, I do not think it was a gift, and I do not think it was a failure. It was a stroke. I rowed competitively anyway, and I hold that as my reading of my own life, not a rule for yours. And we quit every day and start up again the next day. The quitting is honest. So is the coming back.

We do not always get to choose what failure takes from us. We have more influence over what meaning we make of it and what we decide to do next.

Where SportsFlow fits

Reading a failure is a skill, and a skill can be practiced. Two parts of SportsFlow, the platform I build at sportsflow.ai, are built for that. The first is The Setback, a module in development for the stretch after an injury, a loss or a failure. It works from a set of readings over time, because a single reading means little.

The second is the library. Most of the feelings in the table above are among the nineteen in the Weather System. The Still Water series treats regulation as a process: detect, interrupt, shift, integrate. The second reading works at the two ends of it: detect the feeling and its size, then integrate what it is reporting. None of it goes up on the wall.

Still not enlightened, Orion