Ten to One

Why the fighter with better specifications could lose, and why the kill ratio deserves a closer look.

A wall board reads ten to one above a row of ten small jets and one lone jet set apart. At a table a researcher in a cardigan and round glasses lifts a folder holding a single jet out of an open blue archive box, beside the blue words Soviet archives.

At Quantico in 1989, retired Colonel John Boyd told about ten people why most people had the wrong idea about two airplanes.

The MiG-15 could out-climb and out-accelerate the F-86 Sabre. Those advantages made it look like the better fighter. Boyd pointed instead to what the pilot could see and how quickly the plane could respond.

The Sabre had a bubble canopy. Its hydraulic controls let the pilot move the stick "just like power steering in a car." The MiG took longer to change maneuvers.

OODA stands for observe, orient, decide, act. Viewed through that loop, Boyd said, the Sabre was the better airplane. A pilot who could see more and change direction faster had an advantage the specifications missed.

That explanation is easy to trust. It gives you a reason why the machine with better figures might lose. It also gives you something useful to change: how quickly you notice and respond.

I accepted that explanation and repeated the number taught with it: ten to one.

That number seemed to settle the question. The faster loop had won, and the score showed how much.

The number behind it

The U.S. Air Force claimed 792 MiG kills. The ratio taught alongside the loop was about ten to one.

Researchers who compared the claims with Soviet and Chinese records found a much smaller advantage. Their estimates put the whole-war ratio between about 1.1 and 1.8 to one. Sources differ on how far it fell.

The ratio also changed with the person flying the MiG. Doug Dildy put it at about 1.4 to one against Soviet veterans in 1951, and about nine to one against other pilots.

Stephen Robinson, who wrote a 2021 book on Boyd, reads that split as a sign of how much training and experience mattered. Boyd also credited training in the same lecture.

Dick Becker flew in 1951. Soviet loss records confirm his victories. His recollection was blunt: "There was no 14-to-1 kill ratio when I was there."

Even the loop's origin was less tidy than the familiar story. In a margin note, Boyd traced it to his YF-16 and YF-17 flight-test work, rather than Korea.

That note records Boyd's view. It doesn't settle the history. But it makes Korea harder to treat as both the source of the idea and its decisive proof.

Who counted it

A gun camera could show bullets hitting a MiG without showing it fall. Pilots could receive kill credits from that film even when nobody saw the aircraft go down.

Some MiGs also crashed without being shot. Spins and high-speed stalls could destroy them. A loss alone couldn't establish that a Sabre pilot had won by seeing or acting faster.

The scoring loop could still close. Film showed hits. Hits earned credit. Credits raised the total, and the total became evidence for the doctrine.

Credits raised the total, and the total became evidence for the doctrine.

Credits raised the total, and the total became evidence for the doctrine. The film strip, The stamp, The tally board, The doctrine, The blue loop.

Click a numbered dot to see what it stands for. Start with 1.

Once taught, that evidence gave readers another reason to trust the number.

William Lind, a follower of Boyd, repeated the ratio in 1985; Osinga repeated it in his study of Boyd's strategy. Robinson's book gave the ratio, and he later acknowledged that Michael Hankins had corrected his error.

There were earlier doubts. But Thomas McKelvey Cleaver describes the ratio as uncontradicted by information from the other side for forty years. Soviet archives became available after the USSR dissolved.

A stricter film standard would still leave the Air Force counting its own claims. The check that worked came from outside.

Where the loop ends

None of that means the Sabre failed. It held air superiority. Nor does it make the OODA loop useless. Robinson still calls it valuable, and Hankins notes that fighter pilots continue to use it.

Korea's scoring loop and forty-year-closed Soviet archives show a fast loop on a noisy signal can beat a slow loop on a clean one. Waiting for certainty has a cost, too.

But speed doesn't remove the need to ask where the loop ends. A kill credit can end the scoring process while leaving the aircraft's fate unknown. The next decision may arrive quickly without being better informed.

I know that gap from marketing reports I handed over myself. The report could be accurate about the channel while leaving the business outcome unchecked. Handing the report over ended my part, and what the customer got was left for someone else to find. A market-side builder owns an outcome all the way to the customer. Now I check the business outcome with the customer before I call the work done, and I ship the fix myself.

The report could be accurate about the channel while leaving the business outcome unchecked.

The report could be accurate about the channel while leaving the business outcome unchecked. The report, The marketer, The taped-shut box, The customer.

Click a numbered dot to see what it stands for. Start with 1.

That gap between the story a company tells and the pattern it actually runs is shadow strategy. Shadow strategy is the pattern a company follows, whether or not anyone planned it. The stated aim can be learning while the daily process keeps producing a score.

A loop that ends at someone who must answer, as on Toyota's line, has somewhere to take the problem. On Toyota's line, a worker who finds a problem can stop production until someone answers for it. A record the scorer cannot edit gives the score somewhere to meet resistance.

A loop that ends at someone who must answer has somewhere to take the problem.

A loop that ends at someone who must answer has somewhere to take the problem. The worker, The faulty part, The pull cord, The person at the desk.

Click a numbered dot to see what it stands for. Start with 1.

For one number you repeat, who scores it, and how long before someone who gains nothing from it can check?

Power steering makes it easier to turn. It doesn't tell you when you're wrong.

I write about the numbers we trust, and who gets to check them.

Part of The Version That Runs