May 27th, 2006, 9:02 am
..and as a complement to Tibbar's answer, it is not that surprising you should have to do this given it is the same procedure as has to be applied for non-trivial deterministic ODEs and PDEs.However, for the simple polynomial cases you mention, the guess can be highly educated indeed. The reason why you guess 1/3 (W_t)^3 in your example is that this would be the result in the deterministic case, and you know that the Ito result ends up as the deterministic result alongside a correction term due to the non-negligible square term in the expansion. Applying Ito's Lemma to the deterministic guess will make that correction term explicit.I hope this helps.