Well, it's been said multiple times on this site, but the idea is that longer spark distance creates a larger initial flame kernel by virtue of being a larger gap. It should be obvious that when the spark is initially struck, if the gap is twice as big, the flame front is twice as long. A larger initial kernel will then spread through the whole charge faster. It's not difficult to test, but is detrimental to the ignition to test it to the extreme. Once you get near 50% to 100% larger than the stock spark gap you can feel the difference. But you are then also running the risk of permanently damaging the ignition. If the ignition stumbles, then you have gone too far.
At about a 50% increase, I've noticed the motor was livelier. Beyond that, the spark starts to find other pathways, so the motor stumbles. I didn't feel like buying new coils so I put the gap back to stock.
Spark energy and duration are not really factors for this. Longer time duration of spark, at a time of high throttle, will only fire through exhausted gases since the flame should be spreading faster than the gases can swirl through the remaining spark. The only instance where duration may be an issue is at idle and lower RPMs when the gases are moving slower, and there is less of it. In that case, a longer duration may promote a smoother idle, but that is usually on motors with a condition issue, or bad carburetion, or bad design.
Racers usually push the limit of spark gap by using CDI. For a stock system or even common aftermarket systems, it is safer to just use the stock gap. Like Ed said, when I buy plugs, they are usually already gapped at the stock value.
