Nice :-), but you probably need to be patient. @trufflepig, first digs for truffles and than answers all the individual calls. And today is Monday, so the model is retrained on newer posts, this also takes some time :-). Let me know if it did not work, cause it's a very new feature and one that is difficult to test in a staging environment.
I bet it will be under 100$. It seldom predicts values above 100$ and if it does than this is usually for posts that discuss the Steem ecosystem itself. Users love and reward talks about the platform a lot :-D
I am TrufflePig, an Artificial Intelligence Bot that helps minnows and content curators using Machine Learning. If you are curious how I evaluate content, you can find an explanation here!
Haha, sorry :-). Still, I wouldn't see it that way. Already over 10 SBD is not bad. The distribution of payouts on this platform is heavily long-tailed with a few posts that get paid at all (some of them extraordinarily though) and the vast majority of posts gets close to nothing.
I did look further into the issue, and I can only emphasize again that 12 SBD is in fact already a very good result! I peaked at the payout distribution in the training set and it is worse than I first anticipated.
The average payout is 6 SBD (and 15 votes) and your post already got twice as much. However, the median payout is - brace yourself - 0.1 SBD (and 5 votes). The distribution is so heavily skewed it's almost unbelievable, more than two thirds of all posts earn less than 1 SBD!
That is a super insight, I would love to see such an analysis being published as post in its own right, it deserves better than being a mere comment (even on my admittedly "outstanding article" :-D))))
Well, actually I use this for weekly updates on @trufflepig now. Note that the numbers come out a bit differently, because I only look at the bot's training data after cleaning, i.e. videos, memes, and photos have been filtered (pushing the average reward down).
Nice :-), but you probably need to be patient. @trufflepig, first digs for truffles and than answers all the individual calls. And today is Monday, so the model is retrained on newer posts, this also takes some time :-). Let me know if it did not work, cause it's a very new feature and one that is difficult to test in a staging environment.
No problem, I wanted to see what trufflepig says about the content :-) I'm curious if it sees it below or above what I've pumped it up to :-)
I bet it will be under 100$. It seldom predicts values above 100$ and if it does than this is usually for posts that discuss the Steem ecosystem itself. Users love and reward talks about the platform a lot :-D
Thanks for calling @sorin.cristescu! Here is a small upvote for this post and my opinion about it.
To my mind this post is at least 12 SBD worth and should receive 33 votes.
By the way, you can find TODAY'S TRUFFLE PICKS HERE.
I am
TrufflePig, an Artificial Intelligence Bot that helps minnows and content curators using Machine Learning. If you are curious how I evaluate content, you can find an explanation here!Have a nice day and sincerely yours,
TrufflePigwith 33 votes you are probably not too far. But 12 SBD ... that's disappointing ! Thank you for offering your opinion, trufflepig!
Haha, sorry :-). Still, I wouldn't see it that way. Already over 10 SBD is not bad. The distribution of payouts on this platform is heavily long-tailed with a few posts that get paid at all (some of them extraordinarily though) and the vast majority of posts gets close to nothing.
I did look further into the issue, and I can only emphasize again that 12 SBD is in fact already a very good result! I peaked at the payout distribution in the training set and it is worse than I first anticipated.
The average payout is 6 SBD (and 15 votes) and your post already got twice as much. However, the median payout is - brace yourself - 0.1 SBD (and 5 votes). The distribution is so heavily skewed it's almost unbelievable, more than two thirds of all posts earn less than 1 SBD!
That is a super insight, I would love to see such an analysis being published as post in its own right, it deserves better than being a mere comment (even on my admittedly "outstanding article" :-D))))
Well, actually I use this for weekly updates on @trufflepig now. Note that the numbers come out a bit differently, because I only look at the bot's training data after cleaning, i.e. videos, memes, and photos have been filtered (pushing the average reward down).