Friday, 25 July 2014

Column 21, 2014 – The science of catching

Printed in The Cricket Paper issue 93, Friday July 25, 2014.
[Full text below]

Wednesday night I was under a massive skied top edge which went so far up that air traffic control got involved. I had what seemed a good minute while it flirted with the upper atmosphere to debate the relative merits of the English way (fingers pointing down) or the Australian way (fingers pointing up).

Exactly what goes on when we catch a ball is a question that has taxed brighter minds than mine for generations.

The first actions on a flying ball are how and where it is hit: force, direction, elevation. Then there’s gravity, which, all things being equal, should mean it describes a perfect parabola before returning to earth. But things are never equal. Next is air resistance or friction, which will vary greatly depending on the rate and direction of spin, the condition of the ball and the turbulence in the boundary layer around it created by this combination, then wind speed, temperature, altitude and barometric pressure.

Evolutionary biologist Richard Dawkins: “When a man catches a ball he behaves as if he had solved a set of differential equations predicting its trajectory. At some subconscious level, something functionally equivalent to the mathematical calculation is going on.”

The mathematical theories have grand names: ‘Trajectory Projection’, ‘Linear Optical Trajectory’ and ‘Optical Acceleration Cancellation’. Combining most of them to a greater or lesser extent is something called the gaze heuristic. A heuristic is an experience-based problem solving technique – learning by trial and error: we know roughly how a cricket ball will behave in the air because we’ve seen it before. We keep it central in our field of vision using our three-dimensional depth perception to manage relative position: forwards, backwards, left and right, to keep the ball in our crosshairs until we intercept it.

Researchers at EPFL, Switzerland’s federal institute of technology, are using a version of the gaze heuristic to teach a robotic arm to catch objects in under five-hundredths of a second. Real time calculation takes far too long, so the arm uses information gathered from previous similar trajectories, matched to motion-capture studies of the way humans move their hands and fingers to catch. Rather than real time trajectory computation it relies on data from previous experiments. Which is another way of saying experience. The arm is still in development but currently has a catching success rate of nearly 70%.

I’d take that any day. Especially under a skier.

I elected in the end to eschew both English and Australian methods, and tried a third way, known technically as ‘the crocodile’, or colloquially as “a complete hash of it.” Still smoking from atmospheric re-entry, the ball ricocheted off the heel of my hand into my eye, leaving me with a splendid shiner with which to advertise my heuristic ineptitude for the next week or so.

All of which, in case it’s unclear, means I dropped a sitter.

- ends 480 words -

No comments:

Post a Comment