How much is karma worth, after all?

by [anonymous]1 min read27th Aug 201121 comments

10

Personal Blog

It's been a couple days since the funding plea, so I thought I'd like to take this chance to compare self-reported donations to short-term karma gains. Naturally, I voted on none of these comments. Note that after posting this, the karma on these posts will almost definitely change; the values here are for 27/8/11 at around 9:00 GMT.

So, the data:

Note: two people (Kaj_Sotala and Rain) reported monthly commitments, but as far as I understand only the yearly pledge is matched, so for the purposes of this informal study I treat them as reporting X*12 USD donations, instead of X/month.

There's not enough data for an honest causal analysis (I tried), but there are a few observations one can make. Intuitively one expects karma to be determined by the donation amount, the duration of time since the posting, and some unknown error.

First observation: the users with the best USD/karma exchange rate made modest contributions early. Nisan came out best, with $6.25/karma — though some of this karma may be due also to the fantastic signal, on their part, that they overcame a rational hazard to make this donation. (Also, EY responded afterward, confounding the karmic flow with his wake.)

In this spirit, we now name "doing the least restrictive, obviously acceptable thing, instead of doing nothing while contemplating alternatives" Nisan's razor, (ニサンの剃刀, perhaps) unless it happens to have a better, previously-existing name.

Second observation: Hyperbolic discounting is alive and well. Those reporting monthly donations have karma below comparable one-shot donations, though both monthly data points did come slightly later than their one-shot counterparts.

Third observation: Large donations are really inefficient at netting karma. pengvado paid $277.48/karma; no one above 1000USD paid less than $50/karma.

Naturally, there's little point to this analysis. If anyone is trying to maximize net karma by donating to SIAI, something is probably wrong with their priorities.

10

21 comments, sorted by Highlighting new comments since Today at 2:12 PM
New Comment

This could be entitled "how much are dollars worth (in karma)".

I had the same impression. It was backwards!

How much is karma worth, after all?

No (fungible) US dollars?

You get karma for comments, including comments signalling altruism. If you want to get karma via saying that you have donated then it becomes a task of maximising the efficiency of your PR exposure. The same applies to other signals, costly or otherwise.

Donating to charity is a less efficient way to gain lesswrong karma than googling for rationality quotes even in terms of time taken to make a donation, neglecting the cost of a donation itself.

Donating to charity is a less efficient way to gain lesswrong karma than googling for rationality quotes even in terms of time taken to make a donation, neglecting the cost of a donation itself.

I earned more than 100 karma from a single rationality quote, and a max of 56 from any single donation comment.

[-][anonymous]10y 0

No (fungible) US dollars?

Hmmm, I suppose all I would need is a botnet and a bitcoin address to change this. That's somewhat beside the point, though.

The economic language here ("exchange rate", "paying", etc.) is not meant in a serious sense: obviously the donors were not making a donation for the sake of gathering karma. But yet, they did signal, and they did (in a non-serious sense) purchase that signal.

Donating to charity is a less efficient way to gain lesswrong karma than googling for rationality quotes even in terms of time taken to make a donation, neglecting the cost of a donation itself.

Of course it is. As I wrote,

Naturally, there's little point to this analysis. If anyone is trying to maximize net karma by donating to SIAI, something is probably wrong with their priorities.

Of course it is. As I wrote,

No implication that you didn't get it intended. Just emphasizing the scope of the disconnect if even time-it-takes-to-donate is sufficient. :)

Naturally, there's little point to this analysis.

Next time you feel tempted to do something similar, donate the equivalent prorated hourly rate in USD instead? :)

[-][anonymous]10y 3

Next time you feel tempted to do something similar, donate the equivalent prorated hourly rate in USD instead? :)

Maybe if SIAI could get certified by GiveWell, I'd consider it. As it stands, though, they'll have to make do with my free time.

they're just doing it wrong. The way to maximize karma by donating is to make a top-level post talking about your donation, and encouraging others to donate. Comments are small potatoes karma wise

[-][anonymous]10y 1

I don't think you're correct. Rare is the top-level post that beats 100 karma; I can do that with ten or so insightful comments that take much less time to compose.

100 upvotes for a top-level post is 1000 karma, not 100 — upvotes for top-level posts are worth ten times more karma than upvotes for discussion and comments. This makes posts disproportionate sources of karma, even given the greater effort involved in writing them.

Note that this is only true in the Main section, not the Discussion section.

[-][anonymous]10y 2

Oh, is that so? I didn't know that... hunh. Interesting...

This makes posts disproportionate sources of karma, even given the greater effort involved in writing them.

I wouldn't say that with the 'even given' part. It takes a lot of effort to make a good post!

but you can't make those 10 or so comments all about one thing. If you're specifically trying to gain karma from donation, a top level post is the way to go in my book.

Kaj_Solata

Sotala.

Note also that not every comment has had the same amount of time to be upvoted - mine was only posted today, for instance.

[-][anonymous]10y 1

I'm so sorry.

Also, I noted this somewhere, didn't I? Maybe it was in a previous draft.

EDIT: Fixed, on both marks.

No problem. :-)

First thoughts and impressions....

First observation: the users with the best USD/karma exchange rate made modest contributions early.

We have a thread from the last Challenge where the time aspect is roughly equal for all posts. Perhaps the data there will control for this time variable.

Hyperbolic discounting is alive and well.

Potentially not the case; if you allow that some fraction of the upvoters are not aware of the 12-months-matched and are upvoting based on 1-month-matched, the discounting might disappear.

Large donations are really inefficient at netting karma.

This screams karma = log(dollar amount) to me.

[-][anonymous]10y 0

Potentially not the case; if you allow that some fraction of the upvoters are not aware of the 12-months-matched and are upvoting based on 1-month-matched, the discounting might disappear.

While they might not be aware of 12-months-matched, 1000/mo should still rank significantly higher than 2000, and it doesn't.

This screams karma = log(dollar amount) to me.

Me, too; it coheres with what I know about how humans cope with large numbers. I had tried a log regression, but there's just not enough data...