IBM Mainframe Forum

I see a rating of 100% for this post.

I see a rating of 14.29% for some of my posts.

What does it mean? Does 14.29%/100% mean my posts are not helpful? Please explain the meaning of the ratings.

For information, it seems that non-moderators don't get to see the ratings yet anyway. Or at least I don't. Or at least I can't find them. Work-in-progress?

Does 14.29%/100% mean my posts are not helpful?

No Frank, the rating system works proportionally based on the number of THANKS against the number of posts.

I am sorry Frank if it hurts you in anyway. But friends, do we really need a rating system here?

Since it would therefore take a considerable time to become meaningful, I can easily do without it.

Hello,

For my $.02, i believe an automated rating system that is driven off of "thanks" would add very little value to our forum.

But friends, do we really need a rating system here?

I don't believe we need a rating of this sort. It could quickly turn into nothing but a popularity contest with no relationship to the quality/value of the post(s).

I do like the "thanks" feature.

No Frank, the rating system works proportionally based on the number of THANKS against the number of posts.

I am sorry Frank if it hurts you in anyway. But friends, do we really need a rating system here?

I'm not hurt - just surprised! That seems like one of the WORST rating system algorithms ever.

It's nice to be thanked, but that rating system would seem to make it impossible to get a high % unless everybody thanks you every time they post which is unlikely. If that's the best rating system algorithm you have, then I'd rather NOT see any rating system.

BTW, for a good rating algorithm, look at Netflix. They let you rate a move on a scale from 1 to 5 and then average the results. Only the people who bother to rate the movie are included in the results, so there are no negative effects from people who don't bother to rate it. The more people that rate the movie, the more "useful" the ratings become.

I'm not sure the Netflix model would work here directly, because of our target audience.

To run with the film analogy, it would be like taking novice film-makers and having to explain to them how to use the tools of their trade, to achieve the particular effect that they want (or think they want) at that time (fear, suspense, comedy, JCL) and then asking them, still novices who have achieved the effect once but may not really yet understand how to replicate it, to "rate" the quality of the technical advice given. Gets a laugh, 5. Doesn't get a laugh, zero until you get me a laugh, even though I didn't explain the character development to you in the first place.

We can't ask people who don't know enough to do the rating, to do the rating. I think the Thanks is a nice feature to have, but I don't think ratings can work unless it is work we can heap on the moderators, and there doesn't seem to be enough benefit for that to make any sense.

I think the most useful thing would be if we can get that feedback from the Guests. We know there are hundreds hanging on every word we type, but it would be nice to actually know there are some finding solutions here.

Yes, I agree that the quality of the ratings we got (no matter what method we used) would be questionnable. I was just trying to point out what a bad vs good statistical model would be.

These users thanked the author Frank Yaeger for the post:

admin (Fri May 04, 2012 9:17 am)

It's Friday, so I'm going to inject a bit of levity and say that some regulars here would like a button with a different digit sticking up. Now THAT would be an effective measurement!

It certainly would save a lot of typing.