Many options model papers propose an option model, derive a solution, or deal with the framework the model provides in which to price an option, and finally, detail how to calibrate the model to market data. This is equally true of many other derivatives models.

So, what’s the problem with calibration? Consider a number of cars attempting to drive up the summit of a steep mountain. For whatever reason, none of them can decide what gear to drive up the mountain in. They know a number of other cars are trying to drive up the same mountain so they phone around and ask other cars what gear they’re driving up the mountain in and make a decision based on their answer.

The problem is, everyone is driving a different vehicle, and taking a different approach. While a few iterations of phoning around and trying various gears will probably get everyone to agree on a gear, not each gear will be optimised for each car and each approach. The drivers would do better to spend some time determining their own optimal approach, than trying to match the gears used by the other drivers.

More on this in the coming weeks.