Message:

<p>[QUOTE=&quot;RonSanderson, post: 3909653, member: 77413&quot;]The neural nets look for correlations. They really aren’t very bright. But the more data they have the deeper the correlations that can be found.</p><p><br /></p><p>Doug actually mentioned several.</p><ul> <li>All Heritage photos are similar.</li> <li>Older holders may have a lower grade bias.</li> <li>NGC and PCGS and ANACS may have biases.</li> <li>TrueViews are glamour shots.</li> </ul><p>The software may not know any of this they way we do, but if this metadata (TPG, grade, photographer, year graded, coin denomination, year, mint mark) is input with each photo, the correlation will be mathematically derived and weighted.</p><p><br /></p><p>While the experiment may fail, it might just quantity what experience already hints at. If we adjust for the discovered biases we may come out the other side with a repeatable model that has predictive power. Perhaps it won’t give a universal definitive grade, but it might give a prediction of what each TPG would grade it.[/QUOTE]</p><p><br /></p>

Your name or email address:

Do you already have an account?

• No, create an account now.
• Yes, my password is:
• Forgot your password?

Stay logged in