Duly noted. For the record I knew what you meant and feel the same. My apologies if I misrepresented what you said.
I ENJOY this particular discus event, so it is worth throwing it all year long. :)
Probably should be working on the novels a bit more though....
Oh, wow. This is cool. Glad to see I made it in the top 89! I'm not at all surprised I didn't make the cut; I definitely needed more WOW! factor in my item to compete with some of the others, but it is nice to see that I was somewhat on track, particularly given how many times I had to go back to the drawing board with my item. Next year- Top 32!!! ;)
Star Voter Season 6
|
Yep, I'm noticing this as well. Flavor is definitely one of the most important parts for these items, and having a bit of gonzo seems to help a lot as well. I think those help a lot more in capturing the public's vote than having a good balanced item.
I was surprised too that balance seemed to take a back seat this year. It was commented on by judges in the item notes, but several of the top 32 I think are under-priced or have mushy logistical issues.
I am super excited to see my item in the top 89! I must not be not too far off base. :)
| 3 people marked this as a favorite. |
It's because I can train you to be better at the pricing math, and I can train you to be more diligent about how your item works.
I can't train you to be more creative.
(Generic "you," of course.)
| 3 people marked this as a favorite. |
It's because I can train you to be better at the pricing math, and I can train you to be more diligent about how your item works.
I can't train you to be more creative.
(Generic "you," of course.)
And now I really want a Rocky Training Montage with SKR. :(
And now I really want a Rocky Training Montage with SKR. :(It's because I can train you to be better at the pricing math, and I can train you to be more diligent about how your item works.
I can't train you to be more creative.
(Generic "you," of course.)
Now if you have a gripe about an item that is NOT in the top 89, you will have to take that up with your fellow voters, not with the judges.
ALL of the above items had a positive vote count, in other words they did not have more votes against than votes for. In other words, the top 89 comprise ALL items that had a positive vote count, meaning those items that had more votes for than against (and a couple--three or four--that were tied). We felt that was a natural and logical cut off for creating our group of keepers from which we judges then sorted out the top 32. I dont fully understand the math but there is an easy way and a more complicated way to sort it out and we used the more expansive way, meaning more items for us to consider.
First I'd emphasize that I'm quite happy with the net quality of the Top 32 (and 89). On the whole I feel they are the best representation of items that we've had in a RPG Superstar yet.
That said, taking the "net positive votes" items is not in fact a statistically sound method of picking the best items from the voting process. The reason for that is that the algorithm presenting items for voting was not picking items for voting at random, but items within their relative popularity (that is why votes generally were between two good or two poor items). That means that it is readily possible for good items to have a net-negative and poor items to have a net-positive rating.
In summary: the voting system used creates an expansive matrix of relative ratings. Using absolute votes to determine popularity dismisses a tremendous depth of the data available.
The algorithm presenting items for voting was not picking items for voting at random, but items within their relative popularity (that is why votes generally were between two good or two poor items).
Umm, no. I'm fairly certain this is incorrect. Gary or Vic has previously posted to the contrary. Pairings were selected at random, and I recall that Vic posted that a week before voting ended, they'd already gone through over 85% of all possible pairings.
Using absolute votes to determine popularity dismisses a tremendous depth of the data available.
They didn't just use absolute votes. SKR explained the precise method in this post, above.
Short version: They took the Top 200 by the public ranking. But "Top 200" might be meaningless, if some of the ranking is determined by slim, statistically-negligible margins. Within these 200 items, they chose a bar signifying "this score of [up votes - down votes] is what we're considering as our Keep pile" - which is fair, if every one of the 200 items has been compared to (practically) every other one. They chose 0. It's a good choice! -- and, Clark waded through the other 111 just to make sure they didn't miss something brilliant.
Really short version: The top 200 were selected with due consideration towards the full matrix; within the Top 200, everything was checked very thoroughly.
I stand corrected.