There are (very probably around) 1.7x10^11 galaxies in the observable universe. So I don’t understand how can P(Aliens in Milky Way) be so closed to P(Aliens in observable universe)? If P(Aliens in an average galaxy) = 0.0000000001, P(Aliens in observable universe) should be around 1-(1-0.0000000001)^(1.7x10^11)=0.9999999586. I know there are other factors that influence these numbers, but still, even if there’s a only a very slight chance for P(Aliens in Milky Way), then P(Aliens in observable universe) should be almost certain. There are possible rational justifications for the results of this survey, but I think (0.95) most people were victim of a cognitive bias. Scope insensitivity maybe? because 1.7*10^11 galaxies is too big to imagine. What do you think?
Tendency to cooperate on the prisoner’s dilemma was most highly correlated with items in the general leftist political cluster.
I wonder how many people cooperated only (or in part) because they knew the results would be correlated with their (political) views, and they wanted their “tribe”/community/group/etc. to look good. Maybe next year we could say that this result won’t be compared to the other? So if less people cooperate, then it will indicate that maybe some people cooperate for their ‘group’ to look good. But if these people know that I/we want to compare the results we this year in order to verify this hypothesis, they will continue to cooperate. To avoid most of these, we should compare only the people that will have filled the survey for the first time next year. What do you think?
I ended up deleting 40 answers that suggested there were less than ten million or more than eight billion Europeans, on the grounds that people probably weren’t really that far off so it was probably some kind of data entry error, and correcting everyone who entered a reasonable answer in individuals to answer in millions as the question asked.
I think you shouldn’t have corrected anything. When I assign a probability to the correctness of my answer, I included a percentage for having misread the question or made a data entry error.
This year’s results suggest that was no fluke and that we haven’t even learned to overcome the one bias that we can measure super-well and which is most easily trained away. Disappointment!
Would some people be interested in answering 10 such questions and give their confidence about their answer every month? That would provide better statistics and a way to see if we’re improving.
If P(Aliens in an average galaxy) = 0.0000000001, P(Aliens in observable universe) should be around 1-(1-0.0000000001)^(1.7x10^11)=0.9999999586.
Only if our uncertainties about the different galaxies are independent, and don’t depend on a common uncertainty about the laws of nature or something. It’s true that P2>P1, but they can be made arbitrarily close, I think.
One way would be for most of the expectation of aliens to come from expectation that the Fermi Paradox is somehow illusionary. There are probably other ways, but I can’t think of any at the moment.
Toy example: Suppose that your credence in “aliens in an average galaxy” is split across 2 distinct hypotheses: A. Life is very common across the universe, but for some reason we can’t detect it. (with confidence 10^-4) B. Life is not common, but any given galaxy has a 10^-16 chance to develop life. Total confidence that alien life exists in any given galaxy: ~10^-4.
So your confidence in “aliens exist in the observable universe” is likewise split: A. Life is very common across the universe, but for some reason we can’t detect it. (with confidence 10^-4) B. Life is not common, but 1.7*10^11 galaxies means a chance of 1-(1-10^-16)^(1.7*10^11) = ~10^-5 Total confidence that life exists in the observable universe: ~10^-4.
EDIT 3: I retract the following paragraph because I now understand what Wes_W wrote.
I know, that’s why I said “There are possible rational justifications”. I mean your reasoning make sense mathematically. But why would your distribution be two deltas at 10^-4 and 10^-16 and not more continuous? It’s not a rhetorical question, I want to know the answer -if there’s one-, but I don’t see how it could be that way. Do you think you are rationalizing your answer? (again, it’s not a rhetorical question)
EDIT: After reading other comments, I think another way a discontinuity might be justify is like this: going faster than light speed is either possible or not.
A. if it is, then if there’s a sufficiently advance civilisation (anywhere in the Observable Universe) it would probably be able to colonize most of the(ir) observable universe. (so the probability that there are aliens in the Milky Way is similar to the Observable Universe).
B. if it isn’t, then it’s the probability that there are aliens in the Milky Way is a lot lower than in the Observable Universe.
EDIT 2: Can you think of other reasons for the discontinuity? With what probability do you think the speed of light is the maximum speed one can transfer information/energy?
I don’t think I’m rationalizing an answer; I’m not even presenting an answer. I meant only to present a (very simplified) example of how such a conclusion might arise.
I’m totally willing to chalk the survey results up to scale insensitivity, but such results aren’t necessarily nonsensical. It could just mean somebody started with “what credence do I assign that aliens exist and the Fermi Paradox is/isn’t an illusion” and worked backwards from there, rather than pulling a number out of thin air for “chance of life developing in a single galaxy” and then exponentiating.
Since the latter method gives sharply differing results depending on whether you make up a probability a few orders of magnitude above or below 10^-11, I’m not sure working backwards is even a worse idea. At least working backwards won’t give one 99.99999% credence in something merely because their brain is bad at intuitively telling apart 10^-8 and 10^-14.
Edit: I think some degree of dichotomy is plausible here. A lot of intermediate estimates are ruled out by us not seeing aliens everywhere.
Sorry I misunderstood. (Oops) I agree (see my edits in the previous comment). A justify dichotomy is more probable than I initially thought, and probably less people made a scale insensitivity bias than I initially thought.
But why would your distribution be two deltas at 10^-4 and 10^-16 and not more continuous?
Because it’s a toy example and it’s easier to work out the math this way. You can get similar results with more continuous distributions, the math is simply more complicated.
Would some people be interested in answering 10 such questions and give their confidence about their answer every month? That would provide better statistics and a way to see if we’re improving.
EDIT: I just made my first (meta)prediction which is that I’m 50% sure that “I will make good predictions in 2014. (ie. 40 to 60% of my predictions with an estimate between 40 and 60% will be true.)”
There are (very probably around) 1.7x10^11 galaxies in the observable universe. So I don’t understand how can P(Aliens in Milky Way) be so closed to P(Aliens in observable universe)? If P(Aliens in an average galaxy) = 0.0000000001, P(Aliens in observable universe) should be around 1-(1-0.0000000001)^(1.7x10^11)=0.9999999586.
Perhaps this is explainable with reference to why the Great Silence / Fermi paradox is so compelling? That even with very low rates of expansion, the universe should be colonized by now if an advanced alien civilization had arisen at any point in the past billion years or so. Hence, if there’s aliens anywhere, then they should well have a presence here too.
Both are wildly impractical (at least, by modern-human-technology standards), but intergalactic is several orders of magnitude more so. The speed of light really isn’t much of an obstacle within a single galaxy; travel at .01c or less is plenty to populate every solar system in “only” a few million years.
It’s believable that a technologically advanced society can cross a galaxy by star hopping and colonization of successive planets, maybe even without generation ships or cryopreservation. E.g. after taking into account relativistic effects, constant acceleration/deceleration at 1g gets us from Earth to Alpha Centauri and back well within a human lifetime. But you can’t star hop between galaxies. There’s nowhere to pick up supplies aside from maybe hydrogen and helium. Even at full lightspeed you need ships that are capable of running for 100,000 years to reach even the nearest galaxy. Is it feasible to build a fire-and-forget colony ship that could survive 10E5 years in space and arrive in working shape? Maybe if you did it with some really robust panspermia or something, and were willing to lose 99% of the the ships you sent out. I.e. just maybe you could transmit biology, but I very much doubt intergalactic civilization is feasible.
You assumption holds if constant acceleration/deceleration at 1g is vastly easier to achieve than generation ships or cryopreservation. If you assume the opposite, then you suddenly can colonize the entire universe, only very-very slowly. :-)
No, not really. Even if generation ships or cryopreservation are easier to achieve than 1g over intragalactic distances, it still doesn’t seem likely that it’s possible to make them work over the 100,000 lightyears minimum between galaxies. To plausibly ship living beings between galaxies you either have to invent science fictional fantasies like Niven’s stasis fields or figure out how to send a lot of seeds very cheaply and accept that you’ll lose pretty much all of them. I’m not sure even that’s possible.
Even if generation ships or cryopreservation are easier to achieve than 1g over intragalactic distances, it still doesn’t seem likely that it’s possible to make them work over the 100,000 lightyears minimum between galaxies.
To me it seems likely that if if you can cryopreserve someone for a 1000 years, you can cryopreserve someone more or less indefinitely.
This discussion is pointless. What seems likely to me or you now has no connection to actual likelihood of the technology.
Entropy is a thing. Keeping a machine running for 10 years without regular maintenance is challenging. 100 years is very hard but within the realm of feasibility. 1000 years might be doable with advanced enough self-repairing technology and access to sufficient fuel. 100,000 years? There’s no way any moving part of any kind is going to keep going for that long. maybe if you can figure out a way to eliminate all moving parts of any kind; but even then I suspect random radiation and micrometeorites might erode any ship beyond hope of recovery. Perhaps there’s little enough of that in the intergalactic void that intergalaxy travel is possible, but I wouldn’t rate it as likely.
To grab another idea from Niven (specifically the Puppeteers), gravity manipulation to get a small traveling solar system would probably work, though it would take an enormous amount of time. I’m not an astrophysicist, but you could get solar wind to keep protecting you from small stray objects and presumably could watch the path ahead to protect yourself from other collisions.
On MIlky Way vs. Observable universe, I would expect a very high correlation between the results of different galaxies. So simple multiplication is misleading.
That said, even with a very high correlation anything over 1% for Milky way should get you to 99+ for universe.
I admit that I did not seriously consider the number of galaxies in the universe, or realize off the cuff that it was that high and give that enough consideration. I estimated a fairly high number for Milky way but gave only 95% to the universe, which was clearly a mistake.
I remember my thought process going something like this: P (Aliens in Milky way) ~0.75 P (Aliens) ~100 P (Answer pulled from anus on basis of half remembered internet facts is remotely correct) ~0,8
So: P (Aliens) P (Anus) ~0,8 P (Milky aliens) P (Anus) ~0,6
I wonder how many people cooperated only (or in part) because they knew the results would be correlated with their (political) views, and they wanted their “tribe”/community/group/etc. to look good.
I don’t think the responses of people here would be so much affected by directly wanting to present their own social group as good. However (false) correlation between those two could happen just because of framing by other questions.
E.g. the answer to prisoner’s dilemma question might be affected by whether you’ve just answered “I’m associated with the political left” or whether you’ve just answered “I consider rational calculations to be the best way to solve issues”.
If that is the effect causing a false correlation, then adding the statment “these won’t be correlated” woudn’t do any good—in fact, it would only serve as a further activation for the person to enter the political-association frame.
This is a common problem with surveys that isn’t very easy to mitigate. Individually randomizing question order and analyzing differences in correlations based on presented question order helps a bit, but the problem still remains, and the sample size for any such difference-in-correlation analysis becomes increasingly small.
There are (very probably around) 1.7x10^11 galaxies in the observable universe. So I don’t understand how can P(Aliens in Milky Way) be so closed to P(Aliens in observable universe)? If P(Aliens in an average galaxy) = 0.0000000001, P(Aliens in observable universe) should be around 1-(1-0.0000000001)^(1.7x10^11)=0.9999999586. I know there are other factors that influence these numbers, but still, even if there’s a only a very slight chance for P(Aliens in Milky Way), then P(Aliens in observable universe) should be almost certain. There are possible rational justifications for the results of this survey, but I think (0.95) most people were victim of a cognitive bias. Scope insensitivity maybe? because 1.7*10^11 galaxies is too big to imagine. What do you think?
I wonder how many people cooperated only (or in part) because they knew the results would be correlated with their (political) views, and they wanted their “tribe”/community/group/etc. to look good. Maybe next year we could say that this result won’t be compared to the other? So if less people cooperate, then it will indicate that maybe some people cooperate for their ‘group’ to look good. But if these people know that I/we want to compare the results we this year in order to verify this hypothesis, they will continue to cooperate. To avoid most of these, we should compare only the people that will have filled the survey for the first time next year. What do you think?
I think you shouldn’t have corrected anything. When I assign a probability to the correctness of my answer, I included a percentage for having misread the question or made a data entry error.
Would some people be interested in answering 10 such questions and give their confidence about their answer every month? That would provide better statistics and a way to see if we’re improving.
Only if our uncertainties about the different galaxies are independent, and don’t depend on a common uncertainty about the laws of nature or something. It’s true that P2>P1, but they can be made arbitrarily close, I think.
I agree. But I don’t think they can be that strongly dependant (not even close). How could they be?
One way would be for most of the expectation of aliens to come from expectation that the Fermi Paradox is somehow illusionary. There are probably other ways, but I can’t think of any at the moment.
Toy example:
Suppose that your credence in “aliens in an average galaxy” is split across 2 distinct hypotheses:
A. Life is very common across the universe, but for some reason we can’t detect it. (with confidence 10^-4)
B. Life is not common, but any given galaxy has a 10^-16 chance to develop life.
Total confidence that alien life exists in any given galaxy: ~10^-4.
So your confidence in “aliens exist in the observable universe” is likewise split:
A. Life is very common across the universe, but for some reason we can’t detect it. (with confidence 10^-4)
B. Life is not common, but 1.7*10^11 galaxies means a chance of 1-(1-10^-16)^(1.7*10^11) = ~10^-5
Total confidence that life exists in the observable universe: ~10^-4.
EDIT 3: I retract the following paragraph because I now understand what Wes_W wrote.
I know, that’s why I said “There are possible rational justifications”. I mean your reasoning make sense mathematically. But why would your distribution be two deltas at 10^-4 and 10^-16 and not more continuous? It’s not a rhetorical question, I want to know the answer -if there’s one-, but I don’t see how it could be that way. Do you think you are rationalizing your answer? (again, it’s not a rhetorical question)
EDIT: After reading other comments, I think another way a discontinuity might be justify is like this: going faster than light speed is either possible or not.
A. if it is, then if there’s a sufficiently advance civilisation (anywhere in the Observable Universe) it would probably be able to colonize most of the(ir) observable universe. (so the probability that there are aliens in the Milky Way is similar to the Observable Universe).
B. if it isn’t, then it’s the probability that there are aliens in the Milky Way is a lot lower than in the Observable Universe.
EDIT 2: Can you think of other reasons for the discontinuity? With what probability do you think the speed of light is the maximum speed one can transfer information/energy?
I don’t think I’m rationalizing an answer; I’m not even presenting an answer. I meant only to present a (very simplified) example of how such a conclusion might arise.
I’m totally willing to chalk the survey results up to scale insensitivity, but such results aren’t necessarily nonsensical. It could just mean somebody started with “what credence do I assign that aliens exist and the Fermi Paradox is/isn’t an illusion” and worked backwards from there, rather than pulling a number out of thin air for “chance of life developing in a single galaxy” and then exponentiating.
Since the latter method gives sharply differing results depending on whether you make up a probability a few orders of magnitude above or below 10^-11, I’m not sure working backwards is even a worse idea. At least working backwards won’t give one 99.99999% credence in something merely because their brain is bad at intuitively telling apart 10^-8 and 10^-14.
Edit: I think some degree of dichotomy is plausible here. A lot of intermediate estimates are ruled out by us not seeing aliens everywhere.
Sorry I misunderstood. (Oops) I agree (see my edits in the previous comment). A justify dichotomy is more probable than I initially thought, and probably less people made a scale insensitivity bias than I initially thought.
Because it’s a toy example and it’s easier to work out the math this way. You can get similar results with more continuous distributions, the math is simply more complicated.
Ok right. I agree.
There’s two sorts of uncertainty here. The more physical kind: probability that life arises, intelligence evolves, etc etc.
And there’s “our uncertainty” kind of probability—we don’t know what it takes for the life to evolve—and this is common for all galaxies.
There’s both PredictionBook and the Good Judgment Project as venues for this sort of thing.
Thank you.
EDIT: I just made my first (meta)prediction which is that I’m 50% sure that “I will make good predictions in 2014. (ie. 40 to 60% of my predictions with an estimate between 40 and 60% will be true.)”
Perhaps this is explainable with reference to why the Great Silence / Fermi paradox is so compelling? That even with very low rates of expansion, the universe should be colonized by now if an advanced alien civilization had arisen at any point in the past billion years or so. Hence, if there’s aliens anywhere, then they should well have a presence here too.
Intergalactic travel is much harder than intragalactic. It’s conceivable that even civilizations that colonize their galaxy might not make it further.
Why would you think so?
If the speed of light is the limit, both are impractical. If it is not, I don’t see why do you assume that physical distance matters at all.
Both are wildly impractical (at least, by modern-human-technology standards), but intergalactic is several orders of magnitude more so. The speed of light really isn’t much of an obstacle within a single galaxy; travel at .01c or less is plenty to populate every solar system in “only” a few million years.
It’s believable that a technologically advanced society can cross a galaxy by star hopping and colonization of successive planets, maybe even without generation ships or cryopreservation. E.g. after taking into account relativistic effects, constant acceleration/deceleration at 1g gets us from Earth to Alpha Centauri and back well within a human lifetime. But you can’t star hop between galaxies. There’s nowhere to pick up supplies aside from maybe hydrogen and helium. Even at full lightspeed you need ships that are capable of running for 100,000 years to reach even the nearest galaxy. Is it feasible to build a fire-and-forget colony ship that could survive 10E5 years in space and arrive in working shape? Maybe if you did it with some really robust panspermia or something, and were willing to lose 99% of the the ships you sent out. I.e. just maybe you could transmit biology, but I very much doubt intergalactic civilization is feasible.
You assumption holds if constant acceleration/deceleration at 1g is vastly easier to achieve than generation ships or cryopreservation. If you assume the opposite, then you suddenly can colonize the entire universe, only very-very slowly. :-)
No, not really. Even if generation ships or cryopreservation are easier to achieve than 1g over intragalactic distances, it still doesn’t seem likely that it’s possible to make them work over the 100,000 lightyears minimum between galaxies. To plausibly ship living beings between galaxies you either have to invent science fictional fantasies like Niven’s stasis fields or figure out how to send a lot of seeds very cheaply and accept that you’ll lose pretty much all of them. I’m not sure even that’s possible.
To me it seems likely that if if you can cryopreserve someone for a 1000 years, you can cryopreserve someone more or less indefinitely.
This discussion is pointless. What seems likely to me or you now has no connection to actual likelihood of the technology.
Entropy is a thing. Keeping a machine running for 10 years without regular maintenance is challenging. 100 years is very hard but within the realm of feasibility. 1000 years might be doable with advanced enough self-repairing technology and access to sufficient fuel. 100,000 years? There’s no way any moving part of any kind is going to keep going for that long. maybe if you can figure out a way to eliminate all moving parts of any kind; but even then I suspect random radiation and micrometeorites might erode any ship beyond hope of recovery. Perhaps there’s little enough of that in the intergalactic void that intergalaxy travel is possible, but I wouldn’t rate it as likely.
To grab another idea from Niven (specifically the Puppeteers), gravity manipulation to get a small traveling solar system would probably work, though it would take an enormous amount of time. I’m not an astrophysicist, but you could get solar wind to keep protecting you from small stray objects and presumably could watch the path ahead to protect yourself from other collisions.
100,000 years from the perspective of outside observers, the amount of subjective time can be made arbitrarily small.
Yes, but the closer you get to lightspeed the bigger problem you have with any collision with any small particle.
On MIlky Way vs. Observable universe, I would expect a very high correlation between the results of different galaxies. So simple multiplication is misleading.
That said, even with a very high correlation anything over 1% for Milky way should get you to 99+ for universe.
I admit that I did not seriously consider the number of galaxies in the universe, or realize off the cuff that it was that high and give that enough consideration. I estimated a fairly high number for Milky way but gave only 95% to the universe, which was clearly a mistake.
Not necessarily, that depends on the nature of your unsertainty, as Wes_W pointed out elsewhere in the thread.
I remember my thought process going something like this:
P (Aliens in Milky way) ~0.75
P (Aliens) ~100
P (Answer pulled from anus on basis of half remembered internet facts is remotely correct) ~0,8
So:
P (Aliens) P (Anus) ~0,8
P (Milky aliens) P (Anus) ~0,6
I don’t think the responses of people here would be so much affected by directly wanting to present their own social group as good. However (false) correlation between those two could happen just because of framing by other questions.
E.g. the answer to prisoner’s dilemma question might be affected by whether you’ve just answered “I’m associated with the political left” or whether you’ve just answered “I consider rational calculations to be the best way to solve issues”.
If that is the effect causing a false correlation, then adding the statment “these won’t be correlated” woudn’t do any good—in fact, it would only serve as a further activation for the person to enter the political-association frame.
This is a common problem with surveys that isn’t very easy to mitigate. Individually randomizing question order and analyzing differences in correlations based on presented question order helps a bit, but the problem still remains, and the sample size for any such difference-in-correlation analysis becomes increasingly small.