Strange Spheres (extra footage) - Numberphile

Поделиться
HTML-код
  • Опубликовано: 24 ноя 2024

Комментарии •

  • @thomassynths
    @thomassynths 7 лет назад +39

    The main video was great, but this "extras" segment is actually the better video. This video gave me way more intuition than the previous one.

  • @LLoydsensei
    @LLoydsensei 7 лет назад +27

    Very nice video. The actual mathematical definition of the spikiness helped me understand it.

  • @achu11th
    @achu11th 7 лет назад +172

    How many parker circles can one fit in a parker square?

    • @ipadair7345
      @ipadair7345 7 лет назад +7

      A circle with a circumference of Tau*(One radius)+-i.

    • @sebastianespejoloyaga7603
      @sebastianespejoloyaga7603 7 лет назад +20

      A lot, but not enough

    • @achu11th
      @achu11th 7 лет назад +3

      Sebastián Espejo Loyaga so you mean about a parker lot.

    • @AlbySilly
      @AlbySilly 7 лет назад +5

      The answer is yes.

    • @U014B
      @U014B 7 лет назад +2

      Seven.

  • @topilinkala1594
    @topilinkala1594 2 года назад +1

    Remember that one dimensional "ball" is just a line segment and it's radius is half of it's length, so the r/4 bit is 12,5%.

  • @davidgillies620
    @davidgillies620 7 лет назад +8

    The hypervolume of a unit n-ball goes to zero as n goes to infinity. It reaches a maximum for n = 5 (of 8 pi^2/15) and thereafter gets smaller and smaller. Higher dimension topology is really weird.

  • @JouvaMoufette
    @JouvaMoufette 2 года назад +1

    The way I like to look at it is that a set of circles is very optimal at fitting in a square. When add a dimension to make cylinders, and put them in a cube, you have a very similar amount of space that it takes up. But when you turn them into spheres, you have less area the sphere takes up, since you have spots under and above the middle circle that aren't the sphere. So you end up with more space between the spheres than you end up with the circle. And the more dimensions you add, the less space they can fill up, so the more space between them.

  • @badrunna-im
    @badrunna-im 7 лет назад +5

    To get an idea of distances, instead of just taking the radius of the middle sphere, think of the diagonal of an n-dimensional cube. It gets much longer than the side length the higher up you go, but still fits inside the cube.

  • @Turcian
    @Turcian 7 лет назад +21

    Theorem: a Parker Circle can be Parker Squared.
    Corollary: Parker Pi (and hence Parker Tau) are not transcendental numbers.
    Proof: homework exercise for the reader.

    • @andymcl92
      @andymcl92 7 лет назад +2

      Including the line "It is trivial to see that..."

  • @SendyTheEndless
    @SendyTheEndless 7 лет назад +7

    Q: How many sides does a Parker n-gon have
    A: Indeterminate

  • @josephlombardo1246
    @josephlombardo1246 7 лет назад +1

    This explanation was much clearer than in the first video, thank you.

  • @rif6876
    @rif6876 7 лет назад +5

    The first rule of "Playing Fast and Loose" is "Don't worry about it".

    • @natfailsyoutube8163
      @natfailsyoutube8163 7 лет назад +1

      It was a "Parker Explanation"

    • @VikeingBlade
      @VikeingBlade 5 лет назад

      @@natfailsyoutube8163 But a Parker square is a failed attempt at a square, while he didnt even try to explain it (because you shouldnt worry about it)

  • @paulmann1289
    @paulmann1289 7 лет назад +1

    Simple to understand when you think about it: It's bigger on the inside just like a rather famous blue box.

  • @okuno54
    @okuno54 7 лет назад +2

    Oh, now I get the "spiky" spheres thing. It's like spikiness, but without the curvature.

  • @billkillernic
    @billkillernic 7 лет назад +2

    Sphere in greek means ball, Celifos (cell/shell) is only the extreme points that make the outer... shell of the sphere. Football translates in greek to ποδόσφαιρο which word by word in English becomes pedisphere.

  • @KristoffDoe
    @KristoffDoe 7 лет назад +3

    One thing we can say for sure about this box that it is definitely BLUE. :)

  • @recklessroges
    @recklessroges 7 лет назад +1

    A ball is an infinite series of nested spheres; so interchangeable ;-)

  • @joecornfield6329
    @joecornfield6329 2 месяца назад

    The way the spikiness resolves the problem is think about the shape on Automatic for the People by REM. If there were 8 of them in a cube you could fit a central one within them which extends beyond the cube.

  • @VeselkoKelava
    @VeselkoKelava 7 лет назад +5

    What I would like to know is the ratio between the "volumes" of the n dimensional sphere and the box, and how that ratio behaves as you go up in dimensions. Maybe it converges? That would be very interesting.

    • @stevethecatcouch6532
      @stevethecatcouch6532 7 лет назад +2

      I looked at the numbers for the even numbered dimensions up to 32. The volumes of the spheres were outstripping the volumes of the boxes.
      The ratios for even numbered dimensions from 2 to 32 are approximately
      0.308
      0.749
      1.9805
      5.564
      16.335
      49.589
      154.575
      492.294
      1596.137
      5253.993
      17521.223
      59097.318
      201331.945
      692018.789
      2397635.534

    • @VeselkoKelava
      @VeselkoKelava 7 лет назад

      Steve's Mathy Stuff thanks for doing that, those are surprising results

    • @stevethecatcouch6532
      @stevethecatcouch6532 7 лет назад +1

      I just realized that I confused myself by reading too many comments. Some people are talking about 2 x 2 x 2 x ... cubes. I should have made my calculations using a 4 x 4 x 4 x ... cube. I'm off by a factor of 2^k. The actual ratios are
      0.03368825528
      0.0192765711
      0.01170117025
      0.007736233366
      0.005433539366
      0.00398795459
      0.003026657236
      0.002358623777
      0.001877953681
      0.001522195063
      0.001252649559
      0.001044346271
      0.0008806186624
      0.0007500199415
      0.0006444927204
      0.0005582430245

    • @stevethecatcouch6532
      @stevethecatcouch6532 7 лет назад +1

      The formula for the "volume" of the central sphere in 2k-D is (pi^k/k!)*(sqrt(2k) -1)^2k. The "volume" of the cube is 4^2k. So the ratio is [pi^k * (sqrt(2k) -1)^2k]/ [k!*4^2k]. I can't find any clever way to show the limit is 0 as k increases.

    • @redbeam_
      @redbeam_ 6 лет назад

      en.wikipedia.org/wiki/Volume_of_an_n-ball#Low_dimensions

  • @DaanSnqn
    @DaanSnqn 7 лет назад +31

    I wouldn't call it a sphere. I would call it a lemon.

    • @sebastianespejoloyaga7603
      @sebastianespejoloyaga7603 7 лет назад +6

      I wouldn't call it a lemon. I would call it a Parker Circle.

    • @thrownchance
      @thrownchance 7 лет назад +6

      a higher dimensional Parker Sphere

    • @kourii
      @kourii 7 лет назад +2

      +1234Daan4321 is a lemon a parker orange?

    • @DaanSnqn
      @DaanSnqn 7 лет назад

      kourii That can't work you see. Parker oranges are supposed to be blue.

  • @JustOneAsbesto
    @JustOneAsbesto 7 лет назад +5

    Soooo... it's like that thing from the Superman film??

  • @luxembourgcool
    @luxembourgcool 7 лет назад +1

    Because of this function square root of d minus 1,when d is a square number, it is a whole number because square numbers roots are whole numbers and any whole number minus one is a whole number

  • @twwc960
    @twwc960 7 лет назад +15

    Matt and I have very different definitions of the word "spiky".

    • @ryanmuller9497
      @ryanmuller9497 7 лет назад +2

      I'd personally never given much thought to how you'd even mathematically define "spiky", but the definition used by Matt (proportion of shape removed by removing a certain portion of a linear dimension) feels intuitive enough - if you think about a star (a shape most people would consider spiky), you can see that it would give a much lower value than a circle (or even a sphere).
      When I watched the original video, the reason why "spiky" felt reasonable to me as descriptor for higher dimensional shapes is that a sphere in any dimension is based on the surface of points equidistant from some other point, and the conversion from the rectangular coordinates to the distance from the centre is Pythagoras' Theorem, the square root of the sum of squares. The more dimensions you have to split the distance to the centre between, the more extreme points you have where nearly all of the distance is invested into one of the dimensions, while the majority of points that share the distance about evenly between all of the dimensions are even closer to the centre point than usual because there are more dimensions to split the distance between. So, in a circle, there are four extreme points (two for each dimension), and a reasonably large amount of the circle exists near each of those points because there are only two dimensions to split the distance between. In a sphere, there are now six extreme points, and a reduced amount of the sphere exists near each of those points because distance is being split between three dimensions. While higher dimensions are a pain to visualise, and as a purely recreational mathematician I haven't actually studied higher dimensional shapes, it's intuitive enough to me that extrapolating that relationship would create the expectation that as you add more dimensions, and even more extreme points, you'd have less and less of the higher dimensional sphere near the extreme points, which is basically the definition of spiky.

    • @twwc960
      @twwc960 7 лет назад +3

      I still have a couple of problems with the description of "spiky". First, spikiness implies a lack of convexity. Imagine a star shape: if you draw a line segment from a point in one spike to a point in another spike, there are points on the segment outside the star. In other words, the star is not convex. Hyperspheres in any number of dimensions are convex. And how many spikes does a sphere have, anyway? 3? 6? Where are the spikes located? Doesn't the presence of spikes violate the rotational symmetry? When you rotate a sphere, don't the spikes rotate with it? If there are, say, six spikes on a sphere, wouldn't you have to rotate a sphere 90 degrees to rotate one spike into another spike? But a sphere should have continuous rotational symmetry. I'm still having a really hard time visualizing these spikes.

    • @HeavyMetalMouse
      @HeavyMetalMouse 7 лет назад +3

      I felt the same way about the odd description of 'spikiness', since it seemed to imply a lack of convexity, but when he describes the meaning in this video, I get it a little better - perhaps a better description would be they're 'more tightly curved' - since the total curvature along a multi-dimensional 'surface' depends on the curvatures of all the different dimensions of it - a surface of a sphere is 'more tightly curved' for the same radius than a circle is, because there are two directions for it to curve in, and curvature is intuitively multiplicative.
      Now, comparing curvature in 2 dimensions to curvature in 3 dimensions is kind of like comparing apples to oranges, but it is a reasonable way to get a sort of intuitive handle on where the difference comes from. Every point is more tightly curved, but there's a LOT more room for points to exist at.
      The sphere doesn't have some discrete number of 'spikes', it's just that every single point on the sphere is 'sharper' in that sense than the sphere of the next lower dimension.

    • @tiagotiagot
      @tiagotiagot 7 лет назад

      He should've said it is more pointy, not spiky.

    • @twwc960
      @twwc960 7 лет назад +1

      That explanation certainly makes more sense than "spikiness", but if the radius of a hypersphere is 1, the Gaussian curvature will be 1 in any number of dimensions. If it is less than 1, the curvature will be greater than 1 and it will grow with the number of dimensions. If it is greater than 1, the curvature will be less than one, and it will get smaller as the number of dimensions increase.

  • @jazzsoul69
    @jazzsoul69 4 года назад

    Fantastic video

  • @Scripter-bx6zy
    @Scripter-bx6zy 7 лет назад +1

    Could somebody answer this question for me
    So, if I was an 11D creature with an 11D box of side length 4 and an 11D sphere of radius 1, my sphere would not fit into my box?

    • @MagnusSkiptonLLC
      @MagnusSkiptonLLC 7 лет назад +1

      It would fit, the 11D sphere is still only 2 units across in every direction, while the 11D box (I'm assuming all sides are the same length) is 4 units wide, and the hypotenuse across the 10D opening would be ~12.65 units across.

    • @Scripter-bx6zy
      @Scripter-bx6zy 7 лет назад

      I see
      That makes sense
      As much sense as I can make of 11 dimensions, at the very least
      Thank you

  • @Vidley1777
    @Vidley1777 7 лет назад

    Matt is a Parker-Jedi, because he can't controll his Jedi-Power

  • @sunilsrivastava3970
    @sunilsrivastava3970 7 лет назад

    If you take a 2d slice of a 9d (or 10d) structure, going through the centers of the center sphere and at least one of outer spheres, what will it look like ? The center sphere should look like a a circle touching (or going out of) box, yet the outer sphere still has radius 1.

    • @Dexuz
      @Dexuz 2 года назад +1

      4 years later but...
      What is taking a 2D slice out of a 9D object, think about what it would mean to take such a slice out of a 4D hypersphere, it would actually equal to taking a line segment 1D of a 3D object, since the 4D object is not merely composed of lines and faces, but also whole 3D structures.
      (e.g. a cube's outer shell is made out of 6 squares shifted into 3D; a hypercube would be made out of 8 cubes! shifted into 4D, it's outer shell is composed of whole 3D objetcs)
      Therefore, taking out a 2D segment out of a 9D structure, would be comparatively removing nothing, since it is 7 dimensions far from its own, just as 0D (a point, no directions) is merely 3 dimensions far from ours.

  • @eagleontheweb
    @eagleontheweb 7 лет назад

    So a little excel sheet later... Any guesses why the ratio of n-volume of the ball to the n-volume of the cube converges on about .00001 at least as far as the 261st dimension further dimensions are causing overflows .

    • @dlevi67
      @dlevi67 7 лет назад +1

      Limited precision of an 8-byte real...

  • @mansertwo
    @mansertwo 7 лет назад

    that was a parker cube if ive ever seen one

  • @verioffkin
    @verioffkin 7 лет назад

    All these inner points inside sphere, that's what topology primarily cares of, right.

  • @JohnSmith-ch9sm
    @JohnSmith-ch9sm 7 лет назад +15

    I don't understand why this is confusing. The box is no longer a 3-d object, it is an n-dimesional object. The "volume" of the box becomes 4^n. And while I don't know the volume of an n-dimensional hyper sphere off the top of my head it is (something pi)*r^n or for r=1 (something times pi). Subract 4^n-(something times pi) you get a lot of empty space..... I don't know how to visualize it though. :-)

    • @BobStein
      @BobStein 7 лет назад +5

      It's freakier than that. Pi is raised to increasing powers every _other_ dimension. en.wikipedia.org/wiki/Volume_of_an_n-ball#Low_dimensions
      The proportion that a sphere fills a box is only 100% for 1 dimension. It not only decreases, it accelerates from there (the denominator is essentially factorial):
      1 - 100%
      2 - 79% (pi/4)
      3 - 53% (pi/6)
      4 - 31% (pi^2/32)
      5 - 16% (pi^2/60)
      6 - 8% (pi^3/64)
      7 - 4% (pi^3/840)
      8 - 1.6% (pi^4/6144)
      9 - 0.6% (pi^4/15120)
      10 - 0.25% (pi^5/122880)

    • @supersmashbghemming6445
      @supersmashbghemming6445 7 лет назад +1

      It is odd because the radius of the sphere is growing bigger and bigger eventually getting out of the box not that the space that it takes up is getting bigger

    • @Dexuz
      @Dexuz 6 лет назад

      @@supersmashbghemming6445 One year later: The radius is growing bigger and bigger indeed, but the possible directions that the spheres need to cover is growing too, every higher dimension is a new direction so in order for the n-sphere to touch every direction the n-box has, the n-sphere itself is covering less n-volume.
      It's easy to understand in our 3 dimensionality, a sphere covers less volume inside a box than a circle because the 3d box has more directions than the 2d box, our 3d sphere is not getting spikier, its extending to 2 new directions, up and down.
      To understand how geometrical objects work in higher dimensions you analize the sequence of the formulas that rules our dimension in comparison to the lower ones.
      Don't try to use actual third dimensional formulas, they don't make sense in any place other than in our 3d world.

  • @labradoodleandpalz
    @labradoodleandpalz 7 лет назад

    I disagree with that definition of pointy. Imagine I take a spearpoint, then I add some random amount of volume to the back end of the spear.
    This makes the other end less of a percentage of the total volume, so now were suppose to say it's pointier? The pointiness of an object should only be a measure of the given point, (cube vertices are pointier than icosohedron vertices) So the angles around that vertex would be the measure, not anything that measures the entirety of the object.

  • @JesseLH88
    @JesseLH88 7 лет назад

    I'm still confused. If the radius increases without bound, won't the sphere eventually completely contain the box? But how could this be true, since the sphere should be just touching the edges of the corner spheres?

    • @andymcl92
      @andymcl92 7 лет назад +1

      The radius does, but so too does the distance from the centre of the box to the corners. The centre of the box is always distance 2 to the centre of a face, but the corners are sqrt(d*2^2)=sqrt(4d)=2*sqrt(d) away from the centre, so the ball/sphere/round thing (with radius sqrt(d)-1) will always be short of the corners. In fact, as you go to higher dimensions, the corners get further away from the ball (2*sqrt(d)-[sqrt(d)-1] = sqrt(d)+1 ). This means the ball escapes further and further out of the box, but gets further and further away from being totally free!

  • @kezzyhko
    @kezzyhko 7 лет назад

    Is there some formula for how much would you slice off in terms of radius and dimension?

    • @andymcl92
      @andymcl92 7 лет назад

      Probably.

    • @andymcl92
      @andymcl92 7 лет назад +1

      (Yes, there is. Easiest way to work it out is probably with integration in 3D, but I don't have the whiteboard pens for that atm!)

  • @j0nthegreat
    @j0nthegreat 7 лет назад +1

  • @stevethecatcouch6532
    @stevethecatcouch6532 7 лет назад +2

    Has any living mathematician actually played fast and loose? Pascal might have played it, but he wouldn't have called it "fast and loose".

  • @MrMakae90
    @MrMakae90 7 лет назад

    I have a really hard time thinking of the 4 dimension inner space of a hyper-sphere, the 4D space analogous to volume in 3D and area in 2D.

    • @kazedcat
      @kazedcat 7 лет назад

      Lucas Balaminut thinking in 4d is impossible. Understanding it can only be done with lower dimensional slices and shadows. Giving up higher dimensional global understanding makes it easier to slice and dice 4d object to analyze parts in lower dimension. It is also easier two follow the numbers than thinking about it geometrically

  • @rngwrldngnr
    @rngwrldngnr 7 лет назад +2

    Is a circle the 2d equivalent of a sphere or of a ball? And what is the 2d equivalent of the other one called?

    • @WolfgangBernady
      @WolfgangBernady 7 лет назад +10

      A 2D *circle* is the outside/boundary of a 2D *disk* .
      A 3D *sphere* is the outside/boundary of a 3D *ball* .

    • @SendyTheEndless
      @SendyTheEndless 7 лет назад

      A solid circle could be called a disc/disk?

    • @WolfgangBernady
      @WolfgangBernady 7 лет назад +1

      I would rather put the other way round: A disc/disk could be called a solid circle. Since that would start with something which is defined exactly, and then attach a alternative description to it.
      But without further nitpicking: yes, that's right.

    • @natfailsyoutube8163
      @natfailsyoutube8163 7 лет назад +2

      Someone is going to get a maths question on the volume of a sphere and write 0 because boundary has no thickness or volume...

    • @WolfgangBernady
      @WolfgangBernady 7 лет назад +1

      Hmm.. That's a tricky one. Maybe the sphere itself has zero volume, but it still contains a non-zero volume inside?

  • @Dalendrion
    @Dalendrion 7 лет назад

    3:55 Force Slice!

  • @Monothefox
    @Monothefox 7 лет назад

    What if we go down to d=1?

    • @Berniebud
      @Berniebud 7 лет назад +1

      In 1 dimension, there's only one possible shape (a line), so the closet equivalent to 4 circles in a 4x4 box would be 2 2-unit long lines within a 4 unit long segment. That would just mean the two lines would take up the entirety of the box, and there would be no space between them. This actually works with the algorithm described. Square root of 1 - 1 = 0.

  • @gedstrom
    @gedstrom 5 месяцев назад

    What if spacial dimensions beyond 3 don't really exist? Then the whole process becomes meaningless! You might be able to talk about it in terms of algebra, but trying to relate the result back to a physical object becomes meaningless!

  • @linkVIII
    @linkVIII 7 лет назад +2

    so a dyson sphere is a "true" sphere

  • @robertmcbrayer6633
    @robertmcbrayer6633 7 лет назад

    ? what if instead of Nd cube in Nd sphere .we Nd sphere in Nd cube?

  • @zertilus
    @zertilus 7 лет назад

    This whole time I've known about numberphile2 but I've never been subbed! Biggest mistake ever I'll tell ya

  • @alexandergallon8850
    @alexandergallon8850 7 лет назад

    In case anyone was wondering, the volume of the sphere is around 1700 and the volume of the box is around 26000.

    • @ragnkja
      @ragnkja 6 лет назад

      In which number of dimensions? It can't be in 7d because when d=7, 4^d=16348, so 7 is too small, but when d=8, 4^d=65536, so 8 is too big.

  • @achu11th
    @achu11th 7 лет назад +16

    The only issue I have with this explanation is, how can you apply pythagoras to those dimension. Because pythagoras requires the next dimension being 90 degree to all previous dimension in order for it to work. So how does even 4 D work in this situation let alone 9 D.
    My suggestion therefore: not only do the spheres change drastically in their shape but also the square in its volume (by changing its shape). That would allow a 9D cube's size to be 4^9 instead of 4^3 and the sphere in the 9th D won't exceed the 9D cube's volume. A 9 cube would like a 4^3 with somehow a volume of 4^9. Or am I not concentrating carefully enough?
    Thanks for matt parker for bearing the parker memes for more than a year and being understanding with us.

    • @lavaande
      @lavaande 7 лет назад +1

      thats a super nice point!!!!!!!!!

    • @ipadair7345
      @ipadair7345 7 лет назад +55

      achu11th the definition of a dimension is that, each dimension is at one right angle to each other.

    • @achu11th
      @achu11th 7 лет назад +1

      Ipad Air so explain to me how I should imagine this problem in more than 4D (which I can get behind). 9D for me requires a 9 D cube to solve the problem and even then it is hard to imagine it for me. So that's why I asked to begin with. I couldn't see a possibility to make a pythagoras at these high levels of dimensions.

    • @Preedx2
      @Preedx2 7 лет назад +20

      Well, each new direction is separated by 90 degrees from previous ones by definition, that is just how maths works when working with higher dimensions, and that is how mathematicians defined them.
      You could possibly assume something different but then you have to state it clearly so that nobody is confused, and if somebody doesn't do that, you can safely assume that they mean standard definition of a dimension.
      As for second part, yes, with higher dimensions you have more "volume" inside cube, but increase in middle spheres "volume" is not what is so amazing about this demonstration, but the fact that this sphere in higher dimensions has radius simingly bigger than box containing it.
      EDIT: with each new dimension you use the result from pithagoras theorem from previous step, i.e. when you step from 2nd to 3rd dimension you use result of pithagoras in 2nd dimension ( sqrt(2) ) and 1, which gives you sqrt(3).
      Then when you step up to 4th dimension you use sqrt(3) and 1 which gives you sqrt(4).
      Then in 5th dimension you use sqrt(4) and 1, and so on, and so on.

    • @achu11th
      @achu11th 7 лет назад +2

      Preedx2 the second part helped me realise the thinking error, I was thinking about. That's why I said "or I am not concentrating enough?". As for the first part, it is the question of how you can proove it to be an exactly 90 degree angle. I know, that it is mathematical definitions involved saying, that it just is 90 degree each step to another dimension, but it still is a valuable question. What if it wasn't 90 degree dimension hopps and therefore would rely on another formula, which would make the spheres sphere again? Or as I stated, what if the new space allowed you to somehow fix the issue with the packaging? As for the radius getting bigger than the containment, I would say, it is unexplainable for me right now (still keeping it a sphere and inside the box). Maybe it is some sort of relativ distance as in physics, where a meter can be smaller than meter in the blackhole, because of gravitation. Maybe n-dimension, can provide a distorted lenght of what it actually is? Or as I said, maybe pythagoras stops working after 3D.

  • @Gold161803
    @Gold161803 7 лет назад

    If I was being tharra

  • @kanavmehra8022
    @kanavmehra8022 7 лет назад +1

    First

    • @VictorAnsem
      @VictorAnsem 7 лет назад +1

      To eat those oranges with the sticky tape still on :P

  • @joecornfield6329
    @joecornfield6329 2 месяца назад

    The way the spikiness resolves the problem is think about the shape on Automatic for the People by REM. If there were 8 of them in a cube you could fit a central one within them which extends beyond the cube.