Reply to thread

Message: <blockquote data-quote="HardRightEdge" data-source="post: 763480">"True AI" as you describe it, without human parameters, is pure science fiction.Today, IBM's Watson reads images of tumors and diagnoses them, evidently with quite helpful accuracy.  It accomplishes this by having been first fed thousands of images of tumors each identified as having turned out to be cancerous or non-cancerous.  When a new image is presented of a current patient, Watson makes a probabilistic determination of cancer or no cancer. So, If one were to apply the Watson/cancer approach to replays, the first step is to load up the data base with instances of catch vs. no catch.  And there's the rub.  Unlike cancer, there's no objective, measurable way of  determining catch / no catch.  Whatever criteria one uses, there will be disagreement.  The Jesse James no-catch is a good example.  The replay is clear as to what happened.  The rule stipulates the receiver must establish himself as a runner.  Disagreement is intense.  Those who think it should have been a catch want to see a rule change.In answer to the OP question, "can this be fixed?", the answer is "no", at least not to a consensus satisfaction.  No matter how the rule is defined and no matter how clear the replay, the determination is more art than science.</blockquote>

No members online now.

Total: 900 (members: 0, guests: 900)