back to article You wait ages for a neutron star and black hole to collide, then two pairs come along at once

Gravitational wave detectors have reportedly spotted the merger of a black hole and a neutron star – not once but twice in the same month. This is the first time scientists have been able to confirm the detection of the merger of a black hole and neutron star, too. Although several gravitational wave events have been detected …

  1. Phil O'Sophical Silver badge

    Lies, damned lies, and statistics...

    occurs within a billion light years of Earth about once a month

    Given that the signal will take anything up to a billion years to reach us, they can't really claim the event itself 'occurs once a month', though.

    1. Fonant

      Re: Lies, damned lies, and statistics...

      As Einstein observed, the whole concept of time is relative to the observer. The "once a month" is relative to us observing the events on Earth.

      1. Phil O'Sophical Silver badge

        Re: Lies, damned lies, and statistics...

        The "once a month" is relative to us observing the events on Earth.

        Obviously, but the comment in the article says "A merger between a galactic void and the densest type of star probably occurs within a billion light years of Earth about once a month". The events did not happen within a month of each other, they were observed from Earth in that period.

        Observed from somewhere else they might have appeared 1m years apart or more.

    2. oldtaku

      Re: Lies, damned lies, and statistics...

      As long as we see it once a month there's no reason to think the recent past (1Byo is relatively recent) was substantially different from the present. It should still be cranking on at about that rate even if we're only still seeing the old ones.

      If you went back 13By it would be a different story, but not a 'mere' 1By.

      1. Phil O'Sophical Silver badge

        Re: Lies, damned lies, and statistics...

        My point is that one of these happened 900m LY away, and the other was 1bn LY away, so even if they were recorded a few days apart by the time the signals reached us, there was actually 100m years between these two events. If they were both at the same distance then I would agree with you.

        1. Anonymous Coward

          Re: Lies, damned lies, and statistics...

          There was some interval between them. How long that was depends on your rest frame. There's certainly a frame in which they were both simultaneous and the same distance away. Any intertial rest frame are as good as any other.

    3. Anonymous Coward

      Re: Lies, damned lies, and statistics...

      I think you can claim thar, with only fairly mild assumptions.

      First of all if the cosmological principle is true then isotropy and homogeneity will mean that any observer at the current cosmological time will observe about one NS-BH collision within a billion light years a month. It's not just us. So this rate should be approximately constant everywhere.

      However those collisions will happen, generally, at earlier cosmological times. But not very much earlier. So you'd then need to model how the rate of such collisions goes with cosmological time to get an estimate of the number happening at the current cosmological time. I don't know the answer to that: I bet people at LIGO/VIRGO/KAGRA have good estimates because they needed to know how sensitive to make their machines so they needed to do modelling.

      However with some handwaving I can convince myself that it likely won't be declining in the recent past and might even be increasing. Handwaving follows. (Almost) all of the objects that will eventually collide are in galaxies which are gravitationally bound and hence not affected by any expansion, so there's no effect from the expansion of the universe. The early stages of inspirals take a very long time and I would expect that the number of neutron stars and stellar BHs might go up over time as massive stars age and die, so it might even be that the rate of collisions rises over time in the recent past.

      But the last paragraph is a very handwavy argument.

      1. jake Silver badge

        Re: Lies, damned lies, and statistics...

        "But the last paragraph is a very handwavy argument."

        Worthy of a beer, though :-)

  2. sebacoustic

    What about the buses?

    > ..merger events are expected to crop up like buses: none for ages, and then many along at once.

    There is no explanation for this part of the prediction, what am i missing?

    1. Anonymous Coward
      Anonymous Coward

      Re: What about the buses?

      It's random (well, it's obviously not random, but the population is so large that it effectively is random, at least statistically. And random things tend to be a lot more 'lumpy' than people expect, because they expect that if you've just seen some event it's somehow less likely that you'll see another one.

      (There's a rumour which might be true that various 'play songs in random order' algorithms had to be made less random so they appeared 'random' to people, because otherwise they play the same song twice in a row 'too often'.)

      1. Arthur the cat Silver badge

        Re: What about the buses?


    2. Annihilator Silver badge

      Re: What about the buses?

      For the same reason that you wouldn't expect to flip a coin 10 times and see it alternate heads and tails with each flip, in fact that would be rather rare (2 in 2^10, or 1 in 512 - you've got the THT... option and the HTH... option). A somewhat lumpy distribution of heads and tails is far more likely.

      Random distribution is only expected to occur over time.

    3. m4r35n357

      Re: What about the buses?

      I think you are missing that this is a very new field of study, and the detectors are still being perfected. What were people watching on telly 300 years ago, must have been rubbish, because they never wrote about it?

    4. Doctor Syntax Silver badge

      Re: What about the buses?

      We hope buses don't collide that often.

      As to bus bunching, there is an explanation. If a bus is slightly delayed at one point there are, on average, a few more people to board (and subsequently alight) at the remaining stops. Each extra person boarding and alighting takes more time so the bus is more delayed. Those extra people would have been picked up by the next bus so it runs a little faster and catches up with the first bus. Apply in reverse for a bus running early, it catches up with the previous one hence bus companies instruct drivers not to get ahead of schedule.

  3. Pascal Monett Silver badge

    Don't forget, people

    Whatever the event, the black hole always wins.

    1. DJV Silver badge

      Re: Don't forget, people

      That's why you pot it last as it's worth more points (wait, I think I'm getting snooker, pool and Red Dwarf mixed up).

  4. Annihilator Silver badge

    “From these two detections, we can infer that, in a fixed volume of space, there are a few more neutron star-black hole mergers per year than black hole-black hole [ones], but not as many as there are neutron star-neutron star mergers.”

    I'd have expected this, on the assumption that black holes are rarer than neutron stars.

  5. Geoff May (no relation)

    GW200105 and GW200115

    I bet they are going to wish they used the 4 digit year in about 100 years time.

    1. Anonymous Coward

      Re: GW200105 and GW200115

      I thought that. I mean, Y2k was not that long ago, but already people have forgotten, apparently: 'it will always be the 21st century, also 7-bit character codes are enough for anyone ...'. Seriously, what?

      1. teknopaul Silver badge

        Re: GW200105 and GW200115

        Congrats, you managed to express that in 7bit. :)

    2. Anonymous Coward
      Anonymous Coward

      Re: GW200105 and GW200115


      "Gravitational waves are named as GW-year-month-day, so our first observation from 14 September 2015 is GW150914. We realise that this convention suffers from a Y2K-style bug, but by the time we hit 2100, we’ll have so many detections we’ll need a new scheme anyway."

      In much the same way black holes have different naming conventions depending on when/how they were named. Cygnus-X1 being the first x-ray source detected in the constellation Cygnus, some are named after the instrument that detected them (the ones detected by the Rossi X-Ray Timing Explorer satellite are prefixed XTE-xxxxx), M31* is the black hole in the centre of the galaxy M31 which was catalogued by Messier.

      1. Geoff May (no relation)

        Re: GW200105 and GW200115

        Thank you for that, very interesting.

      2. Anonymous Coward

        Re: GW200105 and GW200115

        Yeah, right. So they could with trivial effort have designed a spec as 'GWYYYYMMDD'<tail>? where <tail> may be omitted (that's what the '?' means in my horrid BNF). If <tail> is present it is <number-within-day>?<tag>? where <number-within-day> is '/'<n> and <n> is a positive integer starting from 1, if <tag> is present it is '-'<suitably-constrained-alphanumeric-string>, and <suitably-constrained-alphanumeric-string> comes from some slightly restricted character set (no spaces at least).

        Semantically if <number-within-day> is omitted it defaults to '/1', if <tag> is omitted the <suitably-constrained-alphanumeric-string> defaults to the empty string.

        So that will serve for a thousand years, allows an unbounded number of detections per day (days should be UTC days of course), and allows people to tag detections however they like. The minimal string is GW20150914, two characters longer than the current scheme. GW20150914 is, when canonicalised GW20150914/1. And for instance GW21040103/128-LISA2 is the 128th event detected detected on 3 Jan 2104 and it has a tag which is 'LISA2'. GW21040103 is the first event detected that day.

        To be really precise you'd want to deal with midnight-crossing events and even events which were midnight-crossing at only one instrument. It would be good if <tag> had formalised internal structure so you could express multiple tags.

        I'm not proposing this as a scheme: I am saying that a scheme could easily have been devised which was good for a thousand years and which allowed for an unbounded number of intra-day detections without being more than two-characters more verbose than the current one. There simply is no good excuse not to do that.

        1. Anonymous Coward
          Anonymous Coward

          Re: GW200105 and GW200115

          To extend your logic though, why only 4 digits... that's awfully shortsighted of you and there's simply no excuse. (Incidentally, your system is good for 10,000 years (or just under 8000 really for simplicity), not 1000)

          I'd be fairly confident that in 80 years we understand and can detect these events to such a fidelity that we don't bother, or we catalogue them with a bit more accuracy - ie whether it's a black hole colliding with a black hole, or neutron star etc. And like the original response suggested, there's value in knowing that events with that nomenclature are from a specific period/detection method.

          1. Anonymous Coward

            Re: GW200105 and GW200115

            Yes, you are right, you can make the system for ever with a tiny modification: just assume that the date is of the form <Y><M><D> where <M> & <D> are 2-digit strings, and <Y> is at least 4 digits. So then GW99929281370103/1474372828 was the 1474372828th event from 3 Jan 9992928137. It has problems with BCE events, but those could be fixed.

            Knowing details about the event was what the tag field was for.

            And, surprisingly enough, I can be very sure that GW20150914 was in fact from 14th September 2015 because I can read dates.

            I have huge admiration for the GW detections, but not for this. For the cost of two characters they could have had a system which would compatibly extend indefinitely but instead they chose to make it so that every bit of software written to deal with these things after 2100 will need to have special cases for the early braindead naming scheme. Two characters.

  6. Anonymous Coward
    Anonymous Coward

    How can they maintain a decent S/N ratio?

    There must be a lot of interference from an ever increasing amount people who are dense enough to bend light bumping into each other..


  7. Anonymous Coward
    Anonymous Coward

    Minor things....

    You realize that gravity must follow the same path as light, right? i.e. curved.

    And thus gravity could not be escaping a black hole?

    And that, gravity-isn't-bending-space, is provable?


    Gravity is observed to travel at the speed of light, so between two arbitrary points A and B it travels the same *distance*. Subdivide the path in two, make that midpoint B and it must also travel the same distance to that midway point. Subdivide it to the limit case of distance tends to zero and prove its the same *path*, not just the same distances...

    So gravity is following the same path as light. Since you've observed gravitational lensing, you also know lights path is curved, so gravity's path is also curved.


    If light cannot escape a black hole, neither can gravity, because they follow the same path, everything travelling magic number 'speed of light' follows that same path.... curved.


    And gravity isn't bending space, because gravity's *strength* depends on it's relative *location*, if it bent space, its *location* would depend on its *strength*, ... circular logic, .... so you can see why gravity isn't the cause of the bent space.


    I assume this is all common knowledge?


    That "light-years" thing, (assuming light travels in a straight line and at constant speed, neither of which is true), kinda always bugs me.

    Light doesn't travel in a straight line, gravity doesn't travel in a straight line, when light hits you, it does not carry the information of the curved path it has followed, and you perceive it as if it was a straight path. You have nothing to compare it with, nothing, not even the path of gravity. You perceive light as if it travels in a straight line based on its short distance approximate behaviour, and you project that onto light that's come across a universe as if it travelled in a straight line.

    It is impossible for light to be travelling in a straight line, of course, any non-zero bend in space means the light would travel in a loop and wrap back on itself, and you've already observed gravitational lensing, so you already know its non-zero bend!

    The universe is finite.... everything loops back to the same place eventually, even light. You treat it as if the bending is tiny, and can be ignored, but that's just how you perceive it, because you think light travels in a straight line. Yet you can literally see the edge of the observable's not tiny bending!

    Lights path is not straight.


    Light and matter, they're the same thing, affected the same way. You compare light to matter and get the constant limit case for speed... of course you do: The scale of matter is from its motion, the speed of light is from its motion, they're moving over the same field, so of course if you compare one to the other you get a constant! Both are affected the same way by the same field.

    i.e. the speed is not constant, you just perceive it that way.

    Obvious question: how would you define a straight line in that system? You can no longer use light to define the straight line, because you know it is following a curve and cannot detect that curve! You will always perceive it as if it is straight.

    Well, obviously, a force travelling infinitely fast has zero time to be affected by any bending, so you should 'hypothesize' such a force, attempt to model such a system.

    1. Anonymous Coward

      Re: Minor things....

      Hello anonymous crank. Answers to two of the less mad bits in case anyone else is reading.

      Information does indeed not pass outwards through an event horizon (in classical GR), and that includes gravitational waves. Gravitational waves from collision events originate instead in the deformations of spacetime outside any event horizon.

      And your point about defining a straight line is related to the equivalence principle and in turn to the nature of manifolds which are always 'locally flat' (you need a formal definition of 'local' here which I'm not going to give). Informally, if you have only a single 'straight line' (geodesic) there's not much you can say: you need more than one. For example if you have a pair of straight lines and you observe that they meet at more than one point then you can immediately infer that the spacetime through which they travelled has curvature. And that's exactly what we observe of course: it's gravitational lensing. Where we observe multiple images of a single object, there are families of null geodesics which meet both at the object and in our telescopes.

      I'm not sure if there a gravitationally-lensed objects which are visible to the naked eye: it would be very cool if there are.

      The rest of your comment is, sadly, just the incoherent mumbling of someone who completely fails to understand what they think they understand. There's a term for that.

  8. Aging Hippy

    Software Error?

    Scientists - we've got some odd results and revolutionised science.

    Engineers (otherwise known as cynical bastards) - we've got some odd results, what have you done wrong.

    In 1991 Andrew Lyne of Jodrell Bank claimed to have found the first pulsar planet . Rumour has it that he had borrowed a spreadsheet which had approximated Earth's orbit to be circular.

    1. Anonymous Coward

      Re: Software Error?

      I'm not sure what point you're making. Scientists and engineers are not mutually exclusive (the engineering done at LIGO is deeply amazing for instance). Scientists also check their results very carefully, and when they turn out to be wrong they retract them. The difference is that, in many cases, the things scientists do don't have any practical impact, while in many cases the things engineers do have very practical impact. And they get them wrong sometimes and planes fall out of the sky and people die, for instance.

      In the case of the pulsar-exoplanet story: pulsars are amazingly good places to look for exoplanets as they provide you, for free, with a stupidly-accurate clock attached to the pulsar whose Doppler shift you can then measure over long periods. Lots of people were looking for these in the early 1990s: Andrew Lyne thought he'd found one, announced it, went back (himself and/or his group) and checked the data, realised it was wrong, and retracted it, standing up in front of a bunch of his colleagues at a meeting to do so. That was a brave thing to do and he was rightly lauded for doing so. His group has since gone on to discover the first binary pulsar system, PSR J0737−3039, which is just a wonderful thing, and has provided really good tests of General Relativity.

      A year later, the first exoplanets orbiting a pulsar were, in fact, found, around PSR B1257+12.

      So, well, stop it already.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like

Biting the hand that feeds IT © 1998–2022