It is somewhat puzzling to me that my PredictionBook evangelizing is well received here, but the fraction of LessWrongers that actually use PredictionBook is vanishingly small. Frankly, it is a scandal to Less Wrong that its high-karmamembers don’t bother to publicly record their own predictions and yet continue to expect others to believe in the efficacy of the techniques taught in its core texts, like The Sequences.
If you want us to believe your beliefs pay rent, why not show us the receipts?
So I don’t know about anyone else, but as far as I can tell my own personal true rejection is: It’s just too hard to remember to click over to predictionbook.com and actually type something in when I make a prediction. I’ve tried the things that seem obvious to help with this, but the small inconvenience has so far been too much
PredictionBook is a horrible piece of software that had major features that didn’t even properly work until a couple weeks ago. Is it any surprise it isn’t well-received when it sucks so badly?
The email has worked longer than it has not worked and is, in fact, currently working.
I really don’t think this is correct. The first e-mail I ever received from them was last week. It also sent the exact same e-mail twice. Therefore I still claim that their e-mail system doesn’t work.
In addition to that, the UI is awful, the site is often quite slow, and their statistics package is quite rudimentary. There is no filtering mechanism for determining which predictions are “serious”—the result is that many people post public predictions that should be private, but aren’t.
The first e-mail I ever received from them was last week. It also sent the exact same e-mail twice.
Eh. Gmail collapses the duplicates for me, so I barely noticed. And you are being notified...
(Also, you’ve only used PB since last June or so, while it’s been running since October 2009.)
the UI is awful
It seems pretty straightforward to me.
the site is often quite slow
That was much improved after Trike did the SQL profiling.
and their statistics package is quite rudimentary.
Yes, because varying proper scoring rules are why no one is using it?
There is no filtering mechanism for determining which predictions are “serious”—the result is that many people post public predictions that should be private, but aren’t.
This only affects Happenstance, not recording your own predictions.
All of these are annoying to various extents, but do they really explain the near-zero uptake?
It is somewhat puzzling to me that my PredictionBook evangelizing is well received here
This particular shame based instance of evangelism isn’t well received.
Frankly, it is a scandal to Less Wrong that its high-karma members don’t bother to publicly record their own predictions and yet continue to expect others to believe in the efficacy of the techniques taught in its core texts like The Sequences.
Eliezer frequently makes predictions and even bets. Luke makes predictions from time to time as well. Not sure about Yvain. Your complaint seems to be that they don’t happen to personally use your preferred website.
As far as I’m concerned you would have struggled to have come up with a more powerful way to persuade us to not use prediction book.
Eliezer frequently makes predictions and even bets.
Frequently? The http://wiki.lesswrong.com/wiki/Bets_registry lists just 3 bets by Eliezer (full disclaimer: 1 involving me), which even dating just from 2008 (the first listed one) represents less than 1 a year.
(If we want to bring in the AI box experiments as involving money and so being bets, it’s still less than 1 a year since that pushes the interval back to the early 2000s while only adding in like 4 bets.)
Eliezer frequently makes predictions and even bets. Luke makes predictions from time to time as well. Not sure about Yvain. Your complaint seems to be that they don’t happen to personally use your preferred website.
Maybe I am being too tough on them, but I don’t think so. Yes, Eliezer makes bets now and then; Luke has even used PredictionBook before (he currently has 2 public predictions on his userpage). On the other hand, what would you think of a martial artist who claimed to have techniques superior to those used by the pros (Bayes versus Science), yet refused to spar publicly (let alone fight) more than a few times a year?
As far as I’m concerned you would have struggled to have come up with a more powerful way to persuade us to not use prediction book.
Upon reconsideration, I now see that I was following a poor strategy of increasing PredictionBook usage. I won’t retract my comment, but I probably won’t make one like that again.
On the other hand, what would you think of a martial artist who claimed to have techniques superior to those used by the pros (Bayes versus Science), yet refused to spar publicly (let alone fight) more than a few times a year?
If Luke, Yvain or Eliezer claimed that they were superior at achieving predictionbook status than others and refused to demonstrate then I would see your point. As it happens nothing they have said indicates that they ought to be able to dominate on predictionbook (although I would expect them to be better than average).
I also note that predictionbook represents a lost purpose. If you orient your thinking and what predictions you make according to what will make you most impressive on predicitonbook you will not necessarily think the best thoughts or subject your belief’s actual weak points to testing. This means I’d say it is more useful for those whose status is not tied up with their performance.
Upon reconsideration, I now see that I was following a poor strategy of increasing PredictionBook usage. I won’t retract my comment, but I probably won’t make one like that again.
NEW GAME: name the the equations and diagrams in the photo on the Donate page.
Correct answers win karma, I predict.
For all x, if x is a comment in response to Luke’s ‘NEW GAME’ comment and it contains the correct answer, then x will have positive karma at the prediction closing time.
It turns out that correct answers don’t win as much karma as predictions that correct answers will win karma.
It is somewhat puzzling to me that my PredictionBook evangelizing is well received here, but the fraction of LessWrongers that actually use PredictionBook is vanishingly small. Frankly, it is a scandal to Less Wrong that its high-karma members don’t bother to publicly record their own predictions and yet continue to expect others to believe in the efficacy of the techniques taught in its core texts, like The Sequences.
If you want us to believe your beliefs pay rent, why not show us the receipts?
So I don’t know about anyone else, but as far as I can tell my own personal true rejection is: It’s just too hard to remember to click over to predictionbook.com and actually type something in when I make a prediction. I’ve tried the things that seem obvious to help with this, but the small inconvenience has so far been too much
PredictionBook is a horrible piece of software that had major features that didn’t even properly work until a couple weeks ago. Is it any surprise it isn’t well-received when it sucks so badly?
The email has worked longer than it has not worked and is, in fact, currently working. There were no discernible differences in usage of it...
PB has very consistently not been popular on LW. “Major features not working” is not peoples’ true rejection of it.
I really don’t think this is correct. The first e-mail I ever received from them was last week. It also sent the exact same e-mail twice. Therefore I still claim that their e-mail system doesn’t work.
In addition to that, the UI is awful, the site is often quite slow, and their statistics package is quite rudimentary. There is no filtering mechanism for determining which predictions are “serious”—the result is that many people post public predictions that should be private, but aren’t.
Eh. Gmail collapses the duplicates for me, so I barely noticed. And you are being notified...
(Also, you’ve only used PB since last June or so, while it’s been running since October 2009.)
It seems pretty straightforward to me.
That was much improved after Trike did the SQL profiling.
Yes, because varying proper scoring rules are why no one is using it?
This only affects Happenstance, not recording your own predictions.
All of these are annoying to various extents, but do they really explain the near-zero uptake?
This particular shame based instance of evangelism isn’t well received.
Eliezer frequently makes predictions and even bets. Luke makes predictions from time to time as well. Not sure about Yvain. Your complaint seems to be that they don’t happen to personally use your preferred website.
As far as I’m concerned you would have struggled to have come up with a more powerful way to persuade us to not use prediction book.
Frequently? The http://wiki.lesswrong.com/wiki/Bets_registry lists just 3 bets by Eliezer (full disclaimer: 1 involving me), which even dating just from 2008 (the first listed one) represents less than 1 a year.
(If we want to bring in the AI box experiments as involving money and so being bets, it’s still less than 1 a year since that pushes the interval back to the early 2000s while only adding in like 4 bets.)
Translation from human: “I can think of multiple instances without trying hard.”
Terrorists frequently attack us.
Seems like a reasonable claim. Certainly compared to how often they attack us.
Perhaps. We’ll see.
Maybe I am being too tough on them, but I don’t think so. Yes, Eliezer makes bets now and then; Luke has even used PredictionBook before (he currently has 2 public predictions on his userpage). On the other hand, what would you think of a martial artist who claimed to have techniques superior to those used by the pros (Bayes versus Science), yet refused to spar publicly (let alone fight) more than a few times a year?
Upon reconsideration, I now see that I was following a poor strategy of increasing PredictionBook usage. I won’t retract my comment, but I probably won’t make one like that again.
If Luke, Yvain or Eliezer claimed that they were superior at achieving predictionbook status than others and refused to demonstrate then I would see your point. As it happens nothing they have said indicates that they ought to be able to dominate on predictionbook (although I would expect them to be better than average).
I also note that predictionbook represents a lost purpose. If you orient your thinking and what predictions you make according to what will make you most impressive on predicitonbook you will not necessarily think the best thoughts or subject your belief’s actual weak points to testing. This means I’d say it is more useful for those whose status is not tied up with their performance.
Thankyou.
Bayes’ rule
The possible futures decision tree
Something to do with the normal distribution, top left in blue
AIXI, in red between your hands?
Correct!
Correct!
Not specific enough, I’m afraid.
Yes!
Looks like no-one else is going to answer—what is the thing top left in blue?
If my memory serves, there is a perception-action loop diagram shown quite prominently in it.
Correct! (That specific one was copied from Shane Legg’s dissertation.)