Forums

Having followed the video of the nice man speaking with the English accent, I quickly moved to test my form. I ran the test passes for Pass and Fail, and it worked great!

And then I noticed something odd.

If the ReCaptcha asked me to enter "4445 asdfwerq", and I entered "4445 asdfwer", it would fail. But if I entered "4445 asdfwert" it accepted it! The string was the right length, but the very last character was wrong... and it seemed ok with that! Like I was getting a "Aw, that's close enough, buddy. We'll let it slide!" The form then submits, and continues along as if everything were just spiffy.

Hi PKD,

Max is right - check the Google ReCaptcha site for their docs. As I understand how it works the test mainly uses one of the two words. The second is there to help them transcribe doubtful words in Google Books (or house numbers from Google Maps). So it is possible that 'good enough' is OK with one of the words.

Bob

I looked up Google's reCaptcha, and sure enough, the method employed is: two strings for the user to input - one that Google knows the answer to, and one that they need the answer to. The one they need the answer to is unknown, and is used by them to decipher OCR's. So, as long as your character length is correct, they'll accept "close enough".