RE: Binomial eq..using FindRoot. Bug?

*To*: mathgroup at smc.vnet.net*Subject*: [mg18061] RE: [mg17992] Binomial eq..using FindRoot. Bug?*From*: "J.Guillermo Sanchez" <guillerm at gugu.usal.es>*Date*: Tue, 15 Jun 1999 01:43:22 -0400*Delivery-date*: Tue Jun 15 07:43:19 1999*References*: <s75f9cdf.086@gwia.umsmed.edu>*Sender*: owner-wri-mathgroup at wolfram.com

I post an e-mail for the mathgroup and I have received very nice solution. Thanks every body. But I have a curiosity. My equation was: Binomial[f n, r]/Binomial[n, r] == a. If I resolve using FindRoot in the way, h[n_,f_,a_]:= FindRoot[Binomial[f n, r]/Binomial[n, r] == a , {r, 20}]; h[320,0.95,0.05]. Out: {r->53.3465}, that is OK but if I move Binomial[n, r] to other side (the eq. is the same): g[n_,f_,a_]:= FindRoot[Binomial[f n, r] == a Binomial[n, r] , {r, 20}]; g[320,0.95,0.05]. I recived the message: FindRoot::"cvnwt": "Newton's method failed to converge to the prescribed accuracy after 15 iterations." {r->14.5784} that is wrong. It is a bug or are there any rules to use FindRoot that I didn't apply?.