I have been trying to implement a smile interpolator with polynomial in delta based on the excellent text of Mr Clark. "Foreign exchange option pricing"However, i see some issue with the numbers given for EURUSD 1 year option. Inputs are quoted ATM, market strangle and risky for 1 year: {18.25%,0.95%,-0.6%}Other inputs:Spot 1.3465Df(T) 0.965991736Dd(T) 0.971027976Smile interpolator is as below for T=1fsigma(K)=exp(c0+c1*normsdist(ln(F/k)/exp(co))+c2*(normsdist(ln(F/K)/exp(co))^2The book quotes the following parameter output:c0 -1.701005c1 0.050131c2 0.800801But with these parameters, one should not be able to recover ATM vol of 18.25% since exp(c0)=exp(-1.701005) =0.001825. The other two terms will have to be zero which is not the case. Am i missing something here?