June 29th, 2013, 11:45 am
The BBG estimate sounds perfectly reasonable. In the BS model (and I will leave to you to interpret the abbreviation) the price at option expiry is lognormally distributed under the risk neutral probability measure with mean equal to the forward price. That is not directly helpful to your intuition, since the lognormal distribution is rather skewed. However, the log of the price is normally distributed around the log of the forward price minus one half the term variance. Not knowing anything about the price process for wheat, I'll randomly guess that a 25% volatility per year is reasonable, which would make the mean of the log price equal to ln(730)-0.5*3*0.25^2, which is equal to ln(664.67). The risk neutral probability of ending up in the money is therefore about 49.9%. Throw in three years of discounting at something like 60 bps per year, and the answer is 49%.