Skip to content

Instantly share code, notes, and snippets.

@mrflip
Last active January 12, 2026 08:07
Show Gist options
  • Select an option

  • Save mrflip/a973b1c60f4a38fc3277ddd57ce65b28 to your computer and use it in GitHub Desktop.

Select an option

Save mrflip/a973b1c60f4a38fc3277ddd57ce65b28 to your computer and use it in GitHub Desktop.
Ancient Greek Geometry walkthrough / answers / cheats

Solutions for Ancient Greek Geometry (https://sciencevsmagic.net/geo)

Most solutions taken from the about thread. See the comments below for more additions since my last check-in.

Polygons

Circle Packs

Circumscribed Polygons

Non-Constructible Figures

Abuse of floating-point math can make the widget approve non-constructible polygons (polygons with edge count 7, 9, 11, 13, 14, 18, 19, 21, 22, 23, 25, 26, 27, 28, 29, 31, 33, 35, ..., which cannot be precisely constructed using straightedge and compass):

@Eddy119
Copy link

Eddy119 commented Oct 3, 2025

Star trick optimizations:
9-gon, in-origin, approx, in 27 moves
11-gon, in-origin, approx, in 38 moves
Others here, please feel free to simplify the other ones through the star trick or other methods.

@ILoveMath62
Copy link

ILoveMath62 commented Nov 18, 2025

More heptagon relationships Tetradecagon in 24 moves: https://sciencevsmagic.net/geo/#0A1.1A0.0L1.1L5.5A2.6A1.9A2.0A17.19L18.26A1.N.1A43.43L19.59L0.43L18.18L86.0L81.0L96.96L19.19L102.46L102.102L118.118L5.5L130.130L86.47L140.140L153.153L5.102L110.110L127.115L175.150L18.139L162 Icosahenagon (what I call the henskaicosagon) in 30 moves: https://sciencevsmagic.net/geo/#0A1.1A0.0L1.1L5.5A2.6A1.9A2.0A17.19A2.N.2L27.27L34.1A34.34L3.3L57.2L49.34L35.35L3.3L66.34L50.50L2.2L76.76L5.5L83.66L5.5L90.65L27.27L106.75L51.51L118.49L80.80L129.57L87.87L139.118L136.106L126.90L67.67L183.139L178.83L59.59L216.129L213.216L98.183L113.272L249.229L149.196L161 Icosioctagon in 36 moves: https://sciencevsmagic.net/geo/#0A1.1A0.0L1.1L5.5A2.6A1.9A2.0A17.0L24.N.24L19.1A24.19L40.24L18.18L48.48L5.5L53.40L5.5L58.58L43.43L69.53L28.28L81.81L18.18L95.69L19.19L110.110L55.55L120.95L50.50L129.129L74.74L146.120L64.64L160.146L89.89L179.160L102.102L199.179L127.127L216.199L118.118L232.216L141.141L251.232L157.157L269.251L174.174L294.269L192.192L320.289L0.313L0.320L229.294L213.213L423.229L399.399L266.423L248.396L470.420L446 Tetracontakaidigon or Lifegon (since 42 is the answer to everything) in 50 moves: https://sciencevsmagic.net/geo/#0A1.1A0.0L1.1L5.5A2.6A1.9A2.0A17.N.2L19.19L27.3L18.18L33.1A26.33L2.2L56.2L45.27L3.3L64.3L46.56L5.5L71.64L5.5L78.46L75.75L87.45L68.68L95.71L57.57L110.95L105.105L126.78L48.48L143.87L136.136L161.55L18.18L180.63L19.19L199.110L193.193L220.143L175.175L239.161L235.235L262.126L215.215L285.199L92.92L306.180L84.84L326.239L323.262L346.220L303.285L393.326L154.154L449.306L121.121L477.449L257.257L509.509L370.477L279.279L572.572L417.303L396.396L474.474L635.393L420.420L632.632L669.417L602.602L666.666L704.635L567.567L746.0L739.704L0.505L813.701L536.323L349.349L446.346L373.373L929.370L539.539L966.599L1003

@ILoveMath62
Copy link

ILoveMath62 commented Nov 18, 2025

@mrflip can you change the 17-gon at the top from the 45-move to the current 38-move?* It's still the same.

@ILoveMath62
Copy link

ILoveMath62 commented Nov 18, 2025

When you use the 34-move nonagon, you can replace with the bisector of point 1 above main origin point and point 1 above the previous, then star instead to get 26 moves.

@ILoveMath62
Copy link

ILoveMath62 commented Nov 18, 2025

Using the star trick (on heptagon for example) on the icosanonagon and taking a different tangent, you can get 60 moves.

@Ian-Zander
Copy link

I've been trying to find the 31-gon through brute force, but floating point keeps getting in the way.

@ILoveMath62
Copy link

ILoveMath62 commented Nov 18, 2025

Yeah. Makes sense. Did you think it was even possible to optimize the heptagon?

@ILoveMath62
Copy link

ILoveMath62 commented Nov 18, 2025

I literally joined yesterday.

@Ian-Zander
Copy link

In your 37-gon attempt, move 26 scrambled (weird floating point stuff I don't understand). It basically places multiple lines at once

@Ian-Zander
Copy link

Something I've found while brute forcing is that changing the scale sometimes fixes it

@ILoveMath62
Copy link

ILoveMath62 commented Nov 19, 2025

I don't know where @pizzystrizzy, @bikerusl, @ralpjs, @mrflip, @Doomslug682, or @WaiLFiN went.

@Eddy119
Copy link

Eddy119 commented Nov 20, 2025

Hi, interesting stuff, I am here, I can't look at all of this right now, but I'll look into it I'm the near future, I'm still interested, and wondering how to neusis construct a 11-gon...

Maybe we should make a discord for all of this? It might make discussions easier...

(My discord is the same as my GitHub)

@Ian-Zander
Copy link

A discord sounds cool

@ILoveMath62
Copy link

Hint for faster possible 21-gon

@Ian-Zander
Copy link

Ian-Zander commented Nov 22, 2025

My game isn't accepting it
edit: I was tweaking

@ILoveMath62
Copy link

ILoveMath62 commented Nov 22, 2025

Approximations to be done for more shapes:
43 (approximation used by taking the tangent, please share) Late Nov update: I finished it!
44 (11x4)
45 (5x9)
46 (23x2)
47 (use by tangent, please share) Early Dec update: finished by me!
48 (2⁴x3)
49 (use by tangent, please share) Early Dec update: Finished by me!
50 (25x2)
51 (use triangle on heptadecagon, please share)
52 (13x4)
53 (use tangent, please share) Mid-Dec update: finished by me!
54 (3³x2)
55 (5x11)
56 (7x8)
57 (3x19) Mid-Dec update: finished by me!
58 (29x2)
59 (use tan, please share) New Year's update: finished by me!
60 (3x4x5)
61 (use tan, please share)
62 (2x31)
63 (7x9)
64 (2⁶)
65 (13x5)
66 (2x3x11)
67 (use tan, please share)
68 (4x17)
69 (might be done with trisecting 23-gon, but might fail, please share)
70 (2x5x7)
71 (use tan, please share)
72 (8x9)
73 (pierpont, approx.)
74 (2x37 [please share 37]) Late November update: 37-gon is finished!
75 (3x25)
76 (4x19)
77 (7x11)
78 (2x3x13)
79 (use tan, please share)
80 (2⁴x5)
81 (3⁴ [use tan, please share])
82 (2*41 [please share 41]) Thanksgiving update: 41-gon finished!
83 (use tan, please share)
84 (3x4x7)
85 (5x17 [perfectly constructible as being a multiple of two distinct Fermat primes])
86 (2x43 [please share 43]) Late Nov update: I finished the 43-gon!
87 (3x29)
88 (8x11)
89 (use tan, please share)
90 (2x5x9)
91 (7x13)
92 (4x23)
93 (3x31)
94 (2x47 [please share 47])
95 (5x19)
96 (2⁵x3)
97 (pierpont, approx.)
98 (2x7² [please share 49])
99 (9x11)
100 (4x25)

Italic involves trisecting an angle and Bold can only be made by taking the tangent.

@ILoveMath62
Copy link

Ohhhhh... maybe instead of taking the tangent for the required angle, what if we take the tangent for 2 times the angle? That should give us more shapes.

@Eddy119
Copy link

Eddy119 commented Nov 23, 2025

I suspect approximating polygons with pierpont prime number of sides is probably not that much easier or different than approximating the non-constructible ones that aren't? Pierpont might be only relevant with neusis or other angle trisection, what's your experience? Are Pierpont polygons easier to approximate?

I guess if you had a reliable way to approximate angle trisection it could be easier here...

@ILoveMath62
Copy link

ILoveMath62 commented Nov 23, 2025

9-move Circle Pack 3 alternative

@ILoveMath62
Copy link

ILoveMath62 commented Nov 23, 2025

If you take the 11-move pentagon and remove the pentagon itself, you get the angle for the triacontagon. Using the star trick, this is the only possible 37-move triacontagon. (I don't know why it takes so long to render)

@ILoveMath62
Copy link

ILoveMath62 commented Nov 23, 2025

If you take the 35-move hendecagon (the one I would've shared if I'd never made the 28-move one) and bisect, you get a 46 move icosidigon. I can't share it though because it glitched. This is the link to it before it glitched. Finish the last move to make it.

@ILoveMath62
Copy link

ILoveMath62 commented Nov 23, 2025

From the 75-move nonadecagon, (enneadecagon) I decided to make different circles to get the points required, and actually get the points not on the x-axis, and form the parallel line needed to make the circle required, and use the star trick +several more to get 56 moves.
Edit: I managed to star with the starting 8π/19 angle to save 2 moves from the nonadecagon. New nonadecagon
Edit 2: I made the parallel line in a different way to save 7 moves.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment