I do the same with the visual captchas. I always click less tiles than there should be or, if something looks like it could be confused with the requested objec

I do the same with the visual captchas. I always click less tiles than there should be or, if something looks like it could be confused with the requested object I will click to have it added to the algorithm. With only two of us doing this, I'm sure the same image is served to others and our attempts to muddy the data is stripped out. Maybe tomorrow we'll be counted as three.