I don't think a number system can be inherently wrong any more than gravity can be wrong; but a system can be confusing from a historical perspective. Numbers seems to be highly subject to reanalyses. From that number page, you can get additional examples of Mixe-Zoque numerals. These are the numerals 1-12 in Quetzaltepec Mixe, Juquila Mixe, Chuxnaban Mixe, Tontontepec Mixe, (Highland Popoluca), the latter parenthesized since they switch to Spanish for 6 and higher.
1: tuʼuk, tuʼuk, tuʔuk, toʼk, (tuum)
2: mæhck, mæhck, mææht͡sk, mæhck, (wɨsteen)
3: tɨɡɨɨk, tɨɡɨk, tɨkɨɨk, toohk, (tukuteen)
4: mahktaxk, mahtaxk, taʃk, maktaaxk, (maktasteen)
5: mɨɡɔxk, mɨɡɔxk, mɨkoʃk, muɡɔɔxk, (mosteen)
6: tɨduhk, tɨduhk, tɨtuuhk, tohtɨk
7: wɨxtuhk, jɨxtuhk, hʃtuuhk, wuxtohtɨk
8: tuktuhk, tuktuhk, tuktuuhk, toodohtɨk
9: taxtuhk, taxtuhk, taʃtuuhk, taaxtohɨk
10: mahk, mahk, maahk, mahk
11: mahktuʼuk, mahtuʼuk, maahktuʔuk, maktoʼk
12: mahkmæhck, mahmæhck, mamææht͡sk, makmæhck
Compare to Rayón Zoque, Copainalá Zoque:
1: tumɨ, tumɨ
2: meca, meca
3: tukaʼ, tukaʼy
4: makšku, makškuʼy
5: mosaʼ, mohsaʼy
6: tuhtaʼ, tuhtaʼy
7: kuʼyaʼ, kuʼyaʼy
8: tukuduhtaʼ, tukuthtaʼy
9: makstuhtaʼ, makstuhtaʼy
10: mahkaʼ, mahkaʼy
11: maktumɨ, maktumaʼy
12: makwɨstɨhkaʼ, makwɨstɨhkaʼy
The Zoque numbers for "8" are sufficiently similar to the numbers in the other languages that I could accept that these reflect the earlier situation, whereas "9" diverges substantially.
You can engage in the exercise of comparing the bits of these numerals and discern partial similarities across the languages. Zoque "9" makstuhtaʼ looks like it contains "4" makšku, but it also looks like it contains "10" mahkaʼ ("1 from 10"). Zoque "6" and "3" seem to have common elements. Zoque "12" seems to have the elements of "10" and "2", except that that "2" element itself is only found in Popoluca.
There is a common enough historical pattern across languages that adjacent numbers influence each other. If "8" tukuduhtaʼ looks like "3-6", and if Zoque "9" should have been takstuhta, then it is not surprising that Zoque "4" makšku could have influenced the change of takstuhta to makstuhta. There is an analogous influence in Russian (and maybe all of Slavic, I don't know) where expected nev'at' appears as dev'at', under the influence of des'at' "10", and similarly "8" vos'em was influence by "7" s'em.