The Coding Tournament Experiment
What makes certain programmers better at difficult tasks?
copyright Steve Oakes and Martin Griffin info@vespa.academy
The coding tournament experiment
In the 1980s, two academics, Tom DeMarco and Tim Lister, invited companies to take part in a coding tournament.
Coding tournaments were just getting started back then. Tournament organisers would set a challenge and groups would work against the clock to programme a solution to a problem.
Here’s a picture of one…
copyright Steve Oakes and Martin Griffin info@vespa.academy
copyright Steve Oakes and Martin Griffin info@vespa.academy
The coding tournament experiment
Demarco and Lister’s coding challenge required the following:
Rather than come together in a big hall, software companies had to complete the challenge in their own buildings.
They were each given a maximum time limit.
They had to use their normal offices, and work within business hours.
Two people per company were allowed to enter, but each had to work separately rather than as a pair.
copyright Steve Oakes and Martin Griffin info@vespa.academy
The coding tournament experiment
Each entrant would receive a full report and a scorecard which told them how they’d performed compared to others, allowing companies to see how their programmers were performing relative to their competitors.
Hundreds of software companies entered.
DeMarco and Lister were interested in seeing whether the best programmers were… more experienced, paid more, or spent more time on the project.
copyright Steve Oakes and Martin Griffin info@vespa.academy
The coding tournament experiment
The results were fascinating. Outcomes were widely distributed, with the best scores ten times as high as the worst.
So… did the company teams who did best have the most experienced programmers?
No.
Programmers with 10 years’ experience had the same results as programmers with 2 years’ experience.
copyright Steve Oakes and Martin Griffin info@vespa.academy
The coding tournament experiment
The results were fascinating. Outcomes were widely distributed, with the best scores ten times as high as the worst.
Did the company teams who did best have the programmers on the highest salaries?
No.
Programmers right at the top were paid very slightly more… but there was no strong link.
copyright Steve Oakes and Martin Griffin info@vespa.academy
The coding tournament experiment
The results were fascinating. Outcomes were widely distributed, with the best scores ten times as high as the worst.
Did the company teams who did best take more time to complete the challenge?
No.
The companies producing the best work took slightly less time to complete.
copyright Steve Oakes and Martin Griffin info@vespa.academy
The coding tournament experiment
So what were the factors which led to success?
At first, DeMarco and Lister were puzzled, until they spotted something.
Software programmers entering from the same company had been asked to work solo, not together… but the results from each company were all remarkably similar.
Something about the company programmers worked for, determined whether their work was better or worse.
copyright Steve Oakes and Martin Griffin info@vespa.academy
The coding tournament experiment
So what were the factors which led to success?
So the two researchers sent questionnaires to each company to find out what the successful companies were like. It turned out there was a huge difference between the successful companies and the unsuccessful ones.
Here’s DeMarco and Lister’s results…
copyright Steve Oakes and Martin Griffin info@vespa.academy
The coding tournament experiment
copyright Steve Oakes and Martin Griffin info@vespa.academy
Companies in the top 25%
Companies in the bottom 25%
The coding tournament experiment
copyright Steve Oakes and Martin Griffin info@vespa.academy
Companies in the top 25%
Companies in the bottom 25%
The coding tournament experiment
copyright Steve Oakes and Martin Griffin info@vespa.academy
Companies in the top 25%
Companies in the bottom 25%
The coding tournament experiment
copyright Steve Oakes and Martin Griffin info@vespa.academy
Companies in the top 25%
Companies in the bottom 25%
The coding tournament experiment
copyright Steve Oakes and Martin Griffin info@vespa.academy
Companies in the top 25%
Companies in the bottom 25%
The coding tournament experiment
copyright Steve Oakes and Martin Griffin info@vespa.academy
Companies in the top 25%
Companies in the bottom 25%
The coding tournament experiment
copyright Steve Oakes and Martin Griffin info@vespa.academy
Companies in the top 25%
Companies in the bottom 25%
The coding tournament experiment
Successful programmers worked in environments like this:
copyright Steve Oakes and Martin Griffin info@vespa.academy
Why might this layout encourage success?
The coding tournament experiment
What might this mean for you?
You need both:
Solo study areas where you can concentrate on difficult work, and…
Communal study areas where you can work in pairs or study groups, discussing the most challenging aspects of your subjects.
copyright Steve Oakes and Martin Griffin info@vespa.academy
Audit your environment…
copyright Steve Oakes and Martin Griffin info@vespa.academy
Best solo study areas: | Best communal study areas: |
| |