Zulip Chat Archive

Stream: new members

Topic: How does one show 0 < ![1 / 3, 1 / 3, 1 / 3]?


Janitha Aswedige (May 06 2025 at 19:28):

I'm trying to use the am-gm inequality using "theorem geom_mean_le_arith_mean". And I got

 0 < ![1 / 3, 1 / 3, 1 / 3]

as one of the goals. How do I prove this?

David Renshaw (May 06 2025 at 19:32):

something like ext i; fin_cases i <;> norm_num should work

David Renshaw (May 06 2025 at 19:32):

if there's a better way, I'd be eager to hear about it

Ruben Van de Velde (May 06 2025 at 19:49):

No, that sounds great

Janitha Aswedige (May 06 2025 at 19:57):

ext i gives the following message ; applyExtTheorem only applies to equations, not
0 ≤ ![1 / 3, 1 / 3, 1 / 3] ∧ ¬![1 / 3, 1 / 3, 1 / 3] ≤ 0

Jireh Loreaux (May 06 2025 at 20:02):

I think you want

apply lt_of_strongLT fun i  ?_
fin_cases i
norm_num

(untested though)

David Renshaw (May 06 2025 at 20:08):

import Mathlib

example : 0 < ![(1/3 : ), 1/3, 1/3] := by
  apply lt_of_strongLT fun i  ?_
  fin_cases i <;> norm_num

Last updated: Dec 20 2025 at 21:32 UTC