The Voynich Ninja
A One-Page Ledger Method for Generating Voynich-Like Text - Printable Version

+- The Voynich Ninja (https://www.voynich.ninja)
+-- Forum: Voynich Research (https://www.voynich.ninja/forum-27.html)
+--- Forum: Analysis of the text (https://www.voynich.ninja/forum-41.html)
+--- Thread: A One-Page Ledger Method for Generating Voynich-Like Text (/thread-5752.html)

Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

RE: A One-Page Ledger Method for Generating Voynich-Like Text - bi3mw - 28-05-2026

Here's a very quick preliminary test. In my opinion, the result is acceptable for such a short text.

   

Code:
#Word Types: 2177 #Word Tokens: 9502 #Words top 120 1 123 daiin 2 96 daiinal 3 63 qokainal 4 61 qokeey 5 54 okaiin 6 50 cheol 7 50 chey 8 50 chor 9 49 chol 10 47 chdy 11 47 chhy 12 47 okaiinol 13 42 cheor 14 42 okol 15 39 qokedyol 16 39 sheody 17 38 chedy 18 37 sheedy 19 36 chinal 20 36 qokeedyol 21 35 chear 22 33 cheey 23 32 cheydy 24 32 sheey 25 31 dair 26 31 okedy 27 31 otal 28 31 shdy 29 30 dain 30 29 chdyol 31 29 oteedy 32 29 sheol 33 28 otaiinol 34 28 sheor 35 27 olor 36 27 oteey 37 27 oteeydy 38 27 qokeeydy 39 27 shol 40 26 aiin 41 26 chain 42 26 cheeydy 43 26 okeey 44 26 otey 45 26 qokeydy 46 26 qoky 47 25 chedyol 48 25 choiin 49 25 okeedy 50 24 choldy 51 23 aiiin 52 23 chal 53 23 choky 54 23 okar 55 23 olol 56 23 shedy 57 23 shey 58 22 aiir 59 22 cholky 60 22 okeol 61 22 olky 62 21 chety 63 21 okody 64 21 shdyol 65 21 sheoraiin 66 21 shor 67 21 ykeol 68 20 aiinal 69 20 cheo 70 20 dady 71 20 okeal 72 20 oteal 73 20 qokal 74 20 ytchy 75 20 ytor 76 19 chty 77 19 okey 78 19 oldy 79 19 otar 80 19 otchy 81 19 ykol 82 18 cheedy 83 18 cheeey 84 18 cheoar 85 18 cheoor 86 18 choraiin 87 18 okady 88 18 olar 89 18 olchy 90 18 otear 91 18 oteeey 92 18 oteody 93 18 qoin 94 18 shal 95 18 ykaiin 96 18 ytaiin 97 18 yteol 98 17 chchy 99 17 okor 100 17 otain 101 17 oteol 102 17 oteydy 103 17 qochy 104 17 qokdyol 105 17 qokedy 106 17 qoteo 107 17 sheydy 108 17 shody 109 17 ykeedy 110 16 chir 111 16 choty 112 16 okair 113 16 okeeydy 114 16 olkydy 115 16 otaiin 116 16 qoal 117 16 qokaiin 118 16 qokeeody 119 16 qokor 120 16 qoty


RE: A One-Page Ledger Method for Generating Voynich-Like Text - Dunsel - 28-05-2026

(28-05-2026, 09:58 PM)bi3mw Wrote: You are not allowed to view links. Register or Login to view.Here's a very quick preliminary test. In my opinion, the result is acceptable for such a short text.

And you see what I was saying about word length.  You have no 1, 2 or 3 length words so you distribution is way to the right.  Voynich is word length 5 centered, not 6.

Look at this output from my generator:

Output tokens: 8274
Output vocabulary size: 2525
Output hapax count: 1551
Voynich comparison tokens: 8519
Voynich vocabulary size: 2436
Voynich hapax count: 1642

So, your vocabulary is in the ballpark, maybe a little light.  You add those length 1 and 2 tokens and that'll bring up that vocabulary count and likely bring your word length closer to the Voynich if you keep the token count the same.

Now, look at my top 20 bigrams and trigrams compared to the Voynich.  It still needs some tweaking but it's in the ball park.  Is yours?

Code:
Top 20 output bigrams (100 pages):   ho: 2062   ch: 1689   sh: 1551   ai: 1319   ol: 1241   he: 1196   ey: 1037   in: 989   da: 884   ii: 851   ar: 761   or: 626   th: 621   ct: 611   od: 579   hy: 568   ha: 513   cp: 451   ph: 421   ot: 363 Top 20 Voynich bigrams (100 pages):   ch: 3082   ho: 1920   ai: 1433   in: 1374   ii: 1281   ol: 1239   he: 1116   da: 1116   sh: 1033   dy: 1024   hy: 948   ok: 930   or: 922   qo: 764   ot: 749   ar: 562   ey: 554   od: 548   ee: 541   eo: 467 Top 20 output trigrams (100 pages):   sho: 850   aii: 736   hey: 734   cho: 729   hol: 706   iin: 638   dai: 537   cth: 529   cph: 388   she: 362   che: 361   hod: 307   ckh: 259   oda: 252   ain: 249   hoy: 243   dar: 190   cfh: 188   chy: 187   har: 184 Top 20 Voynich trigrams (100 pages):   cho: 1236   iin: 1181   aii: 1139   dai: 728   che: 718   hol: 596   chy: 492   sho: 467   hor: 467   cth: 435   tch: 420   kch: 392   edy: 366   qok: 359   she: 295   hey: 291   ody: 253   qot: 251   otc: 244   heo: 241

Look at my word length compared to the Voynich then compare it to yours.

   

Look at my zipf curve compared to the Voynich. Does yours compare?  I'm betting that until you fix the word length it'll be a "no".

   

These are the the very basic numbers a generator needs to be "plausible" with before you can even call it a generator.  AND, to be a workable generator it has to be mostly blind about it.  I do use a seed page. After that, it generates the rest on it's own just using my ledger.  And all the numbers you see mine generate are emergent.  I didn't create any code to force a zipf curve.  I do have code that manages word length and that's something a scribe could do with their eyeballs.

The big thing you have to think about.  If you sat down and tried to create something that looked like the Voynich and YOU were in the 15th century, what "tools" would you need other than a quill, some ink and maybe some ponce to erase things with.  Any generator we come up with today needs to be human 15th century doable. Which is why the Naibbe cypher looks great and would technically be human doable but would require dice and cards and lookup tables... extremely cumbersome.  Those are the type things you need to consider.  

Creating Voynich-like text is easy.

Creating a method that explains how a 15th century scribe could do it, that's the hard part.

Edit: If you want to really see how you're doing, put up some real Voynich comparison charts, word length for word length.

RE: A One-Page Ledger Method for Generating Voynich-Like Text - Dunsel - 28-05-2026

Ran some new tests on my generated code vs the Voynich.

This test measures novelty, predictability, word beginnings, word middles, word endings, and how strongly words cluster into families.  It's from some code I developed months ago.  It basically trains on half the text then tests on the other half to see how they compare.   

In this test the generated code is mostly in the ballpark with the exception of bigram reject. That's basically a measure of, look at the bigrams on one half of the text and then see which of those bigrams are in the other half.  In the case of my generator, it's pretty low which means, it's producing too many close relatives. It stays inside the learned families extremely well, but the real herbal text appears to generate more new combinations while still obeying its structural constraints. 

BPC is bits per character which measures entropy or how predictable text is.  BPC2 uses bigrams, BPC3 trigrams.  ΔBPC is the difference between the two.

The ED1 tests look at how many words in the training set are edit distance 1 or copies of words in the test set.  Token asks how many words have a close relative in the test set and type asks how many unique words have a close relative.  Again, it's off by a bit and is probably related to the bigram reject difference.  97% of the unique training words are ED 1 or 0 to a test word.  

And I think I know why these are issues. I tried to keep everything a copy or ED1 from it's source and it almost never generates a random new word.  It appears there's a bit more novelty in the Voynich than I'm allowing for.

So again, not perfect.  In the neighborhood?  Mostly, but needs work.

MetricHerbal VoynichRK42Difference
Bigram reject %0.56550.1044much lower
BPC22.44722.8717higher
BPC32.19782.5697higher
ΔBPC0.24940.3020higher
Initial BPC3.06923.1479slightly higher
Medial BPC3.33063.4473slightly higher
Final BPC2.60142.9747higher
ED≤1 Token95.7988%98.9295%higher
ED≤1 Type88.0416%96.9503%much higher


RE: A One-Page Ledger Method for Generating Voynich-Like Text - bi3mw - 29-05-2026

(28-05-2026, 10:22 PM)Dunsel Wrote: You are not allowed to view links. Register or Login to view.And you see what I was saying about word length.  You have no 1, 2 or 3 length words so you distribution is way to the right.  Voynich is word length 5 centered, not 6.

Problem solved—not perfect, but good enough.

   

============================================================
PAGE 49
============================================================

chedyol ckh otecheo otdy otain qok sheol choeeyky
ytchy ykhol okor cheolky oteydy okar chockhy ytol
oteydy ch oldain sheol daar sheol cheesey char
l olal ykol cholody okodar choolchdy qokshey shain
qokchy otalshy oka qoteshy ok daseky ira otair
oteydy aiirody qooraiin cheky chky oleeol ote shey
qotor aiindy oleey chody qotey kal ytal oldyol
ok qo otdy choty oldyol otockey dair otardar
sheekchey qokchedy sheey choindy qocphy qotey ykchy chteey
oteor rot ykeey qokychdy oteal che chchy qokeeydy
qo ch shcthy cheedalaiin aiolky qokaiiin otal aithar
ykchy okear dair oteor sheey qo aiiinol

============================================================
PAGE 50
============================================================

sheol oka ykeydy oka ykchy chear otey aiir
lol yld choolky aiir chear osh oka choldar
charain qochy cheor otdady ol ok shoky oleydy
qokeey ytal ol ytey shdy shkshy chty daiin
chty okoraiin otor chodaiin ollchy olardy darody yteey
ykar cheor ykeey daiin okedy okar okaiin qofaiin
shkear otosey daiin sheey oleydy okeey qokeey qoeeor
ol checphey qochey chindy ch shdchy cheokeor daiin
chey daiir ykhol oltaiin okol ch shedy shechy
chor che che chedy chtain shedy chekydy daiin
qot oteeo qoin qolchy okchy cheol qot oksheo
okshor shol ykdy qotarain okaiin qolshey cheol

RE: A One-Page Ledger Method for Generating Voynich-Like Text - Dunsel - 29-05-2026

(29-05-2026, 05:16 AM)bi3mw Wrote: You are not allowed to view links. Register or Login to view.
(28-05-2026, 10:22 PM)Dunsel Wrote: You are not allowed to view links. Register or Login to view.And you see what I was saying about word length.  You have no 1, 2 or 3 length words so you distribution is way to the right.  Voynich is word length 5 centered, not 6.

Problem solved—not perfect, but good enough.

Now you're looking much more Voynich and you have at least one metric in the ballpark.  I'm guessing adding those short words reduced all the long words and brought that chart back closer to reality.  You're now matching word length distribution for a natural language.  The Voynich tends to have slightly shorter words than a natural language.  And, that bell shaped standard distribution should also start to bring your zipf curve into line.

RE: A One-Page Ledger Method for Generating Voynich-Like Text - bi3mw - 29-05-2026

(29-05-2026, 10:52 AM)Dunsel Wrote: You are not allowed to view links. Register or Login to view.And, that bell shaped standard distribution should also start to bring your zipf curve into line.

   

RE: A One-Page Ledger Method for Generating Voynich-Like Text - bi3mw - 30-05-2026

In conclusion, I would like to clarify once again that my model is based primarily (87.4%) on a table containing 485 entries (syllable combinations). The ledger words only fill in the gaps not covered by short_words and fixed words. They therefore account for only a small percentage, but they are necessary to ensure that all words are generated. Otherwise, the model as a whole would not function properly.

Proportions of word generation:
-----------------------------------------

Short words:  1634 (17.2%)
Common words:  4474 (47.1%)
Common combinations:  2195 (23.1%)
Ledger words:  1197 (12.6%)
Total: 9500

The high proportion of predefined words suggests that the author may indeed have worked with a tables book or combinatorics. While this would be labour-intensive, it would still be feasible for a 15th-century author.

Of course, that doesn't prove the manuscript is meaningless. A table book could also encode a real language or a cipher system.


.zip   generated.zip (Size: 20.11 KB / Downloads: 8)

RE: A One-Page Ledger Method for Generating Voynich-Like Text - Torsten - 30-05-2026

(26-05-2026, 04:11 PM)Dunsel Wrote: You are not allowed to view links. Register or Login to view.
(26-05-2026, 03:07 PM)Torsten Wrote: You are not allowed to view links. Register or Login to view.I prefer to analyze the text the scribe actually wrote rather than speculate about the texts he didn't write — since that number is endless.

I will politely disagree with this statement. A model does not just have to explain what the scribe wrote. It also has to explain why the text did not fall apart into garbage.  A lot of objections in these discussions are really variations of “why don’t we see this?” or “why didn’t the scribe do that?” They are questions about constraint.

In my own experiments, most copy/mutate systems fail pretty quickly unless the constraints are tight enough. They drift, bloat, repeat badly, or start producing obvious nonsense. And, given enough generations, even my generator has issues. That failure is informative. It tells us the Voynich text is not just “anything goes” mutation. So I think the unwritten text matters too. If a proposed method could easily generate endless bad Voynichese, then we need to explain why the manuscript consistently avoided doing exactly that.

If we don't explore the possibilities of what wasn't created then we'll never have acceptance for what is created.

You raise a fair point — but it's a different question from the one I answered. Zandbergen asked why some unspecified modifications are absent. I said: because the scribe took one path and not another, and we can only analyze the path taken. That stands.

Your question is different: why doesn't the text fall apart into garbage? That's a question about constraint, and it has a straightforward answer.

The copying process inherits structure from the source. Each new word is a modification of an existing word that already has valid structure. A modification of a valid word almost always produces another valid word — because the stroke-level constraints are preserved by the modifications taken. The structure doesn't need to be maintained by external rules — it's carried forward by the copying process itself.

And the scribe has eyes. If a modification produces something that looks wrong, he sees it and doesn't write it. Your own "don't look stupid guardrails" describe this exactly. The scribe isn't following a ledger. He's looking at what he produced and judging whether it looks acceptable. Human aesthetic judgment is the constraint. 

His visual judgment follows the stroke-level structure of the writing system. Schwerdtfeger described in 2008 four design rules: (1) line-glyphs can follow line-glyphs or 'a'; (2) curve-glyphs and 'a' can follow curve-glyphs; (3) the 'l'-glyph can be used as a curve-glyph or as a line-glyph; and (4) gallows glyphs count as curve glyphs (see Timm & Schinner 2020, p. 10). These aren't rules the scribe memorized — they're descriptions of what visual similarity produces. A scribe modifying strokes naturally stays within these constraints because crossing them would look wrong to his eye.

This is also why your generator struggles — and why any generator will. For the human mind it is difficult not to repeat itself. The variation comes from imperfect repetition — not from randomness. The consistency comes from habitual patterns — not from explicit rules. A computer needs both dice (for variation) and rules (for consistency) as separate mechanisms. A human produces both from one source — his own cognitive habits. That's why modeling a human scribe in Python is, as you put it, "a nightmare." You can't model the human mind like a computer — because the human produces variation and structure from the same process, while the computer needs separate mechanisms for each.

The manuscript doesn't fall apart because every word was seen and accepted by a human before it became part of the source pool for the next word. The constraints are the pattern recognition abilities and the habits of the scribe himself.

RE: A One-Page Ledger Method for Generating Voynich-Like Text - Dunsel - 30-05-2026

(30-05-2026, 07:30 PM)Torsten Wrote: You are not allowed to view links. Register or Login to view.The manuscript doesn't fall apart because every word was seen and accepted by a human before it became part of the source pool for the next word. The constraints are the pattern recognition abilities and the habits of the scribe himself.

I couldn't agree with you more. And again, that brings human psychology into the equation. Which patterns?  And then, how do you model those patterns?  At what point do the weights and rules become more than a 15th century scribe can do?  That's where I am now.

For examaple:

In the other post on <ed> I mentioned I did some machine learning, trying to find syllables.  I refined that down to where only 3 words in Takahashi and 14 in Zandbergen/Landini can't be mapped has having good, learnable n-grams in Scribe 1 content.  It came up with a count of 150 main n-grams and 50 more that that act as connectors.  I added that learned chunk data into my generator and it produced the code blocks I pasted below.  

The ledger model still attempts to do a mutation and validates it. The new word is then, it's passed off to the chunk model to see if it contains valid n-gram chunks with valid neighbors.  It looks more Voynich like that previous runs and still matches statistics pretty well. 

But, is this still human doable.  I think so but, it may require more than just the scribe's eyes . There may have been an audible component, where the scribes were familiar enough with the output tokens that they pronounced them.  I know I pronounce Voynich words to see if they sound right. I pronounce words as I write them.  Might a Voynich scribe do the same?

Code:
<f1v> <f1v.P1.1;K>        fchol.shaiin.odad.fchol.odd.ose.cfhealy.chol <f1v.P1.2;K>        chotey.yoo.or.syaiir.cther.ckhea.chtain <f1v.P1.3;K>        pchol.ycheeo.chol.shodd.pose.odad <f1v.P1.4;K>        pydaey.okain.yor.pchol.oydar.pchol.kshiy.yorr <f1v.P1.5;K>        fchol.os.fchol.s.daiin <f1v.P1.6;K>        osdar.okod.qos.daiim.pshol.ycheeo <f1v.P1.7;K>        oar.okeol.oateos.opeos.daiin.fchol.chool.mor <f1v.P1.8;K>        os.chol.oos.pchol.kaiin <f1v.P1.9;K>        cfhoain.fcho.chal.rokain.cchol.fchol.cthar.s <f1v.P1.10;K>        kor.chol.chols.ydain.s.lchol.sshey <f1v.P1.11;K>        chool.diiin.fchol.kor.s.fchol.yocho <f25r> <f25r.P1.1;K>        ros.dloctha.sory.dain.eey.dloctha.koshey.shol.daraii <f25r.P1.2;K>        kshy.dlocta.ycho.kshy.dain.ikan <f25r.P1.3;K>        cthooldar.chodaiils.cthy.okaiirg.chodaiil.ctholdar <f25r.P1.4;K>        chotey.odar.cfhol.csol.shory.dloctha.ecthoary <f25r.P1.5;K>        kdail.dlocta.dlocfha.cthey.dair.dain.daid <f25r.P1.6;K>        cphoy.dlocta.dloctha.tchol.dloctha.cthoary <f25r.P2.7;K>        okar.chal.dair.cphoy.cfal.dlocta.okaiir.eos <f25r.P2.8;K>        oycheey.cphodaiil.daire.chhealy.chealy <f25r.P2.9;K>        oar.tchol.dlocta.tchol.dlocfha.dlocpha.dlocka.dlocpha.cphealy <f25r.P2.10;K>      kaiirg.eckhoary.fchol.tchol.dlocphe.dair <f25r.P2.11;K>      kail.dlocka.os.pchil.otchol.dlocka.ckdey.pchol.dsory <f25r.P2.12;K>      dlocpha.pchol.cthol.oaky.chodaiila.ctoil <f50v> <f50v.P1.1;K>        dan.cphoy.chor.dshkoldy.dan <f50v.P1.2;K>        kyey.teody.ydain.cphar.dkchar.y <f50v.P1.3;K>        cfhol.cphesaiin.chor.dshpoldy.sckhey <f50v.P1.4;K>        cshyds.kodshey.cphoy.cthar.chor.kon.choydr.dain <f50v.P1.5;K>        oas.chol.cphar.shok.shon.otol <f50v.P1.6;K>        shos.rchol.oiin.scphey.ooiin.sho.cphar.oydar <f50v.P2.7;K>        y.ro.oeykoal.rchol.kchol.fchol.chodor <f50v.P2.8;K>        rol.olys.ytais.eolys.dshpoldy.cthey.y.sho <f50v.P2.9;K>        okol.fchol.dshkoldy.y.dshpoldy <f50v.P2.10;K>      cfhyl.chor.ytain.stchey.chsy <f50v.P2.11;K>      eoe.ckhar.rchol.dshpolday.ddchar.oteos <f50v.P2.12;K>      cphey.ckhey.oekoal.oteol.chey.stchey.cfhol <f50v.P2.13;K>      y.oas.chey.chol.alg <f50v.P2.14;K>      pydeeyl.rol Page-count rule: output pages = Voynich pages = 100 Voynich pages available: 225 Voynich pages compared: 100 Generated pages: 100 Output tokens: 8234 Output vocabulary size: 2463 Output hapax count: 1557 Voynich comparison tokens: 8519 Voynich vocabulary size: 2436 Voynich hapax count: 1642 Ledger invalid: 0 Operation counts:   external_copy: 1065   external_copy_fallback: 14   external_ed1: 1961   external_short_copy: 13   initial_gallows_construct: 88   initial_gallows_seed: 10   local_copy: 951   local_copy_fallback: 164   local_ed1: 1954   novel_ed1: 1830   seed: 184 Gallows action counts:   attested_ed1: 2903   gallows_delete: 54   gallows_insert: 63   gallows_substitute_in: 35   gallows_substitute_out: 121   gallows_swap: 162   initial_construct: 88   initial_seed: 10 Top 20 output bigrams (100 pages):   ho: 2214   ch: 1934   he: 1295   sh: 1274   ol: 1254   ai: 1116   ey: 997   in: 887   da: 815   th: 778   ct: 749   or: 739   ii: 720   ar: 685   ha: 510   od: 499   hy: 468   cp: 382   eo: 373   ee: 368 Top 20 Voynich bigrams (100 pages):   ch: 3082   ho: 1920   ai: 1433   in: 1374   ii: 1281   ol: 1239   he: 1116   da: 1116   sh: 1033   dy: 1024   hy: 948   ok: 930   or: 922   qo: 764   ot: 749   ar: 562   ey: 554   od: 548   ee: 541   eo: 467 Top 20 output trigrams (100 pages):   cho: 969   hol: 836   sho: 702   cth: 682   hey: 677   aii: 571   dai: 520   iin: 505   che: 497   cph: 352   ain: 333   she: 332   tho: 275   hod: 262   oda: 226   kch: 221   ckh: 218   eey: 204   cfh: 186   dar: 181 Top 20 Voynich trigrams (100 pages):   cho: 1236   iin: 1181   aii: 1139   dai: 728   che: 718   hol: 596   chy: 492   sho: 467   hor: 467   cth: 435   tch: 420   kch: 392   edy: 366   qok: 359   she: 295   hey: 291   ody: 253   qot: 251   otc: 244   heo: 241

After generating 1000 pages, it starts to look less Voynich-like but still has familiar patterns.

Code:
<f500v> <f500v.P1.1;K>      shokcheey.ods.lochy.ol.dar.cfhaiin.shkaiir.moaokchy <f500v.P1.2;K>      opsho.cfhol.kooishey.ol.cfhaiin.ol <f500v.P1.3;K>      oyr.maokchy.kachys.ool.chol.oal.chol.shoasho <f500v.P2.4;K>      fool.oal.keishey.oalg.cthol.lochy.y.ol <f500v.P2.5;K>      maokhy.cfhol.dain.ool.oal.cpod <f500v.P2.6;K>      oaokchy.ol.ool.ol.oaikdon <f500v.P2.7;K>      cphey.y.shear.ord.ckhaiin.shtasho.cthey.ckhhaeiin.oaidoon <f500v.P2.8;K>      ool.maokhg.okchey.okeey.sor.cpod.lol <f500v.P2.9;K>      old.deydd.cthol.she.dain <f500v.P2.10;K>      ol.shey.dair.kol.ols.ostairin.tol.roe.ol <f500v.P2.11;K>      sheo.old.cfhol.shory.ol.cod.shyoolso.ror.cfhoaiin <f500v.P2.12;K>      maokhy.ois.shoolsho.lshoolsho.oil.toche.skaiir

And here's page 10,000.  I assumed if the model was going to collapse, it would have done so by this point.  I can't provide any stats to compare to Voynich other than to say it's a bit heavy on longer words which could be adjusted.

Code:
<f5000v> <f5000v.P1.1;K>      kore.dain.shey.chy.otaiin.qosoier.kore <f5000v.P1.2;K>      raiisoey.cthal.cthodaiils.pdeiimy.airg.qosoilr.ylsoier <f5000v.P1.3;K>      dain.chal.y.chal.oor.chodain.cthodaiils.odaiin <f5000v.P1.4;K>      dpdaeim.cthy.ore.qosoier.daicthy.y.kshy <f5000v.P1.5;K>      qosoier.odar.dain.kshy.olsdeoey.eefy <f5000v.P1.6;K>      chy.dain.oor.qoosoieor.dair.lshytoo.cthodaiils.qore <f5000v.P1.7;K>      dain.airg.cphey.kodaiin.dair.alshyoe <f5000v.P2.8;K>      fairg.s.sshyood.qosoilr.chok <f5000v.P2.9;K>      sory.olsoior.olsooer.dair.alshyey.alshiey <f5000v.P2.10;K>    fair.cthodaiils.fairg.olsoier.oor.sshysod.dair.pachys.s <f5000v.P2.11;K>    oir.alshyoe.alshyee.qosoier.shey.chok.olsoar.ckhealy


RE: A One-Page Ledger Method for Generating Voynich-Like Text - Dunsel - 31-05-2026

(30-05-2026, 04:28 PM)bi3mw Wrote: You are not allowed to view links. Register or Login to view.Of course, that doesn't prove the manuscript is meaningless. A table book could also encode a real language or a cipher system.

Ran some of my tests on your output.  Hope you don't mind a critical analysis.  

TestHis OutputTakahashi SliceComment
Tokens9,4479,447Matched sample size
Types1,7572,739Far fewer unique words
Hapax6831,856Far fewer one-off words
Type/Token Ratio18.6%29.0%More repetitive vocabulary
Hapax / Type Share38.9%67.8%Major novelty deficit
Mean Word Length5.044.87Very close
Zipf Slope-1.021-0.841Too steep. Common words dominate too much
Length Distribution JSD0.0047-Excellent match
ED≤1 Token Hit89.55%-Lower than expected
ED≤1 Type Hit77.52%-Weak word-family structure
Bigram Reject Rate2.94%-Reasonably good
Trigram Overlap79.84%-Moderate match

My eyeballs say your text looks good, almost too good perhaps. The Voynich has it's own license on weird. Which I suspect is why your hapax and novelty are low.