Thank you @shaqmeister, your observations about the word “DOTT” and the possibility that he may have indicated “IUBILAE” through the terminal characters at the end of the lines were particularly enlightening.

Now I need to summarize everything and also carry out that Codex verification I mentioned to you.

Posted by: @lendor-77

↑

Another aspect, although probably a marginal one, is that words such as “DROGA” and “SANGUE” do not seem especially consistent with his usual themes or imagery, whereas terms like “UCCIDERE” and “MORIRE” appear much more in line with them.

Although not necessarily so in the context they appear, doesn’t ‘DROGA’ have the meaning ‘drug’ (“I LIKE KILLING BECAUSE…”) and ‘SANGUE’ ‘blood’?

Bonjour @shaqmeister! Sorry I disappeared for a while, but I’ve been busy working on another anonymous letter that I think may be connected to the Monster of Florence, possibly even linked to the murder of a female taxi driver.

Going back to this discussion, I noticed that you also thought about Zodiac’s line: “I LIKE KILLING PEOPLE BECAUSE IT IS SO MUCH FUN…”. I have the impression that the Monster was more drawn to terms related to killing and violence rather than references to blood or drugs. Just a personal observation — maybe a bit too speculative.

Anyway, while rewatching some old livestreams analyzing this envelope, I found some interesting connections between Zodiac’s victims and those of the Monster of Florence. For example, I noticed that 51 could correspond to the total obtained by adding the number of victims Zodiac claimed up to 1974 — 37 — and the 14 victims of the Monster of Florence between 1974 and 1985.

37 + 14 = 51.

Another curious detail: if you count the characters by separating them according to punctuation marks (counting punctuation as well), you get:

• “DOTT.” = 5 elements
• “DELLA MONICA SILVIA PROCURA DELLA REPUBLI-” = 37 elements
• “CA 50100 FIRENZE” = 14 elements

Once again, the numbers 37 and 14 appear.

Sorry, I forgot to mention that number 51 is the issue number of the magazine Gente used for the clipping letters.

Posted by: @lendor-77

↑

Bonjour @shaqmeister! Sorry I disappeared for a while

Ciao, @lendor-77.

That’s OK. I was, in any case, assuming that you were busy working on your planned Codex verification and didn’t want to disturb you while you were doing that. I had also come to a point, myself, where I felt I didn’t have much else to contribute at this time. That might change with your ongoing work, however!

Hola @shaqmeister!
Sorry it took me longer than expected, but unfortunately I’m the kind of person who tends to get lost easily. 😅

Over the past week I got completely stuck on a letter sent to P. Vagheggi and, in the end, I believe I managed to find (in my opinion) a possible interpretative key for part of one of the Monster’s most complex letters. The juiciest part, however, is still unknown to me, and that is precisely why I think it makes sense to publish it here on the forum: it is literally driving me crazy.

Coming back to the point, I reached a conclusion while trying to verify whether the passage:

IUBILAE ⇒ SLAVES

could represent a genuinely meaningful structure or simply a coincidence that can be easily reproduced.

The analysis is not limited to the single case, but instead builds a general verification method. The envelope sequence is treated as an ordered source of characters from which possible Latin words or stems compatible with the original order of the letters can be extracted. This is therefore not a free anagram: characters may be skipped, but it is not possible to go backwards within the sequence, except for minimal local swaps between adjacent letters.

The transcription used derives directly from the address written on the envelope:

“DOTT DELLA MONICA SILVIA PROCURA DELLA REPUBLICA 50I00 FIRENZE”

with numerical conversion according to the A=0 system:

• 00 ⇒ A
• 0 ⇒ A
• 5 ⇒ F

Within this framework the sequence emerges:

IUBLIAE ⇒ IUBILAE

obtained through a small local swap between adjacent letters. The A derives from the numerical block “00”, an element considered important because it introduces a non-arbitrary internal constraint.

Once the Latin stem IUBILAE has been identified, the procedure applies all possible alphabetical shifts (+1 up to +25). With shift +10, the sequence produces:

SELSVAO

From here, one residual letter at a time is removed in order to verify whether the remaining six letters can form English words through anagramming. Among the various outputs appears:

SLAVES + O

where the residual letter O is converted through A=0 into the number:

O = 14

I have however pointed out that SLAVES is not the only possible output: the same procedure also generates words such as:

• SALVES
• SOLVES
• SOAVES
• LOAVES
• SALVOS

For this reason, the value of the result is not placed in the existence of the single word SLAVES alone, but rather in the simultaneous combination of multiple factors:

• independent rule for letter selection;
• anomalous concentration of single-contribution pages in the final part of the envelope;
• formation of a recognizable Latin stem;
• presence of the internal constraint given by the “00” block;
• production of a semantically relevant English word;
• presence of a coherent numerical residual.

In order to verify whether the phenomenon was common or merely random, I then constructed a broader statistical analysis using:

• a Latin dictionary containing 130,451 entries;
• an English dictionary containing 48,192 entries.

Two search fields were defined:

• narrow field: complete 7-letter Latin words;
• broad field: initial 7-letter stems extracted from longer words.

The analysis shows that the simple passage:

Latin stem ⇒ shift ⇒ English word + residual

is not extremely rare when applied to thousands of words. A significant percentage of sequences indeed produce at least one English solution after alphabetical shifts.

For this reason, I introduced a second phase of comparative control through alternative clusters, constructed according to different rules derived from the same envelope:

• pages whose digit sum equals 10;
• pages containing the number 7;
• cluster 42–47;
• pages ending in 5;
• last lines of the envelope;
• clusters with digit sums 8 or 9;
• single pages.

Each cluster was subjected to the same procedure: Latin search, alphabetical shift, English verification, and direct checking of possible words.

The overall result is that almost all alternative clusters produce:

• null results;
• isolated outputs;
• or weak and poorly relevant coincidences.

None replicate the same combined structure found in the IUBILAE case.

The conclusion of the work is therefore cautious: the method does not independently demonstrate the existence of an intentional encrypted message, but it does identify a non-trivial combination of conditions that appears more specific than the ordinary lexical coincidences produced by broader tests.

Hi @lendor-77.

It looks like you have been very busy with your investigations and, as can happen, having to fight off distractions at the same time.

It is clear you have put together a very comprehensive study here, and it is certainly interesting seeing your results. I very much look forward to reading your full analysis, when you are ready to post that, and will certainly study it carefully.

A quick question for you, @lendor-77.

I’m wanting to make sure that I fully understand your “second phase of comparative control through alternative clusters” and am getting myself a little confused as to what you’re actually doing here. I’m sure it will become clear with your full presentation, but in the meanwhile, could you clarify a couple of terms and other points for me?

Just generally, to begin with, does your 7-point list here describe seven steps in a single control example (i.e., all taken together), or are we talking about 7 separate control examples?

Also, I’m not sure I’m clear about what some of the terms mean here as, for example:

• “pages containing the number 7.” Pages of what? The edition of Gente? The latin dictionary? Are we talking page numbers, or if the number appears anywhere on the page?
• what exactly am I to understand by “digit sum?” and finally
• what am I to understand is meant when you refer to “clusters?” What, precisely, are your calling a cluster?

I am prepared to admit that I sometimes have mental blocks like this, so forgive me if I am missing the obvious. 🤔 

Here’s the analysis!

Good questions, and no worries at all, I probably explained that section too compactly. 😅

What I call “clusters” are simply different groups of letters extracted from the envelope according to specific selection rules. The idea was to test whether results similar to the IUBILAE case could also emerge from other independently selected subsets of the envelope.

So, for example:

• “pages containing the number 7” means pages from the envelope reconstruction whose page number contains the digit 7 (such as 37, 47, 57, 78);
• “digit sum” means the sum of the digits of a page number:
  • page 37 => 3 + 7 = 10
  • page 45 => 4 + 5 = 9
• “pages with final digit 5” simply means pages ending in 5;
• “pages with initial digit 2” means pages beginning with 2.

The word “cluster” is therefore just my shorthand for:
“a subset of letters extracted according to a recognizable rule.”

Each cluster is then tested independently using the same method:

• search for compatible Latin words/stems;
• apply alphabetical shifts;
• check for possible English outputs plus residuals.

The goal was basically comparative: if many unrelated clusters produced structures similar to IUBILAE ⇒ SLAVES, then the main case would become much weaker statistically.

So the controls were not designed to “prove” the main result, but rather to estimate how frequently comparable structures emerge elsewhere in the same envelope.

Very clear now, @lendor-77. Thank you for answering so quickly. I will read the full analysis and then come back after. 

Once again, lendor, very clearly presented and well-argued. I imagine it must have taken some time just to set up the tests in Codex.

As I think I’ve noted before, I am nothing even close to being an able statistician and, in fact, always found statistics the most mysterious part of mathematics. Consequently, I can only give a generalised, tentative response from my own initial reading of these results.

Certainly, I can easily see that everything is sound, both in its reasoning and the conclusion you draw, up to the first result. I am, however, finding myself feeling less persuaded by the method applied in the second part, looking for “alternative clusters” matching some or other “recognizable criteria.”

I think the main reason I feel uneasy, from this point on, is that we are no longer faced with a complete analysis, as in the first instance where all matching words from the Latin dictionary were considered, and where all those transposed words from the English dictionary (plus residual) were taken into account. In contrast, however, in the second part we are merely selecting alternative clusters according to only 9 sets of criteria, randomly selected, and then stopping there.

Don’t we, likewise, have to test all possible alternative clusters before we could reach a meaningful result? And how would we even go about selecting all such criteria? What would that even mean?

Hi @shaqmeister! First of all, thank you for taking the time to look at this analysis. Technically, it did not actually take me that long, also because, as I mentioned, at some point I got a bit distracted by other things. Paradoxically, the most demanding part was putting the report together in Italian and then translating it into understandable English.

But yes, you are absolutely right: nine clusters are very few and certainly cannot be considered exhaustive. My idea, however, was to start from clusters that had at least some logical coherence or minimal recognisability, rather than generating completely arbitrary combinations.

To overcome this limitation, a broader and more systematic analysis could be set up. For example, we could ask Codex to automatically search all possible clusters within a certain maximum length, say up to 10 letters and with a minimum length of 7, and then verify whether within those clusters it is possible to obtain a 7-letter Latin word consecutive within the envelope sequence, while respecting the constraints we discussed: at most one simple anagram (that is, a swap between two letters), a fixed-key shift, and naturally the final presence of a meaningful English word.

What do you think?

Buongiorno, @lendor-77.

So, from my first thoughts, I would be looking at your proposal, and your analysis as a whole, like this…

As to this last suggestion, I don’t think that we would be able to reach any result that we didn’t have already. The initial “broad field” analysis, for example, pulled out all the viable candidates meeting your criteria and with no constraints on the cluster lengths analysed. The candidates produced from this method ultimately all had 7 character length. Adding the constraint that the cluster lengths should not be greater than 10, whilst again including all such clusters, would then simply produce the same result set, would it not?

For a further consideration, let us determine the total number of possible combinations of 7 characters, forming anything whatsoever, from the original 53 (is it?) character set. (And, for a first approximation, let us rule out any character switching.) If I’m not in error, there are approximately 1.54 x 10^8 such combinations. Then, from the “broad field” analysis, we have, say, 4 x 10^3 matching stems, which corresponds to around 1 in every 3.8 x 10^4, or roughly 0.003%.

If we are next assuming that there is only this one candidate combination that is constrained to letters from single pages of Gente, then the probability of a candidate cluster meeting this additional criterion (single page source) as well as the others would be the same 0.003%, if my reasoning is correct.

In fact, perhaps I could simplify the approach by avoiding entirely the problem of “manually selected” clusters.

Let me see if I understood correctly: take directly all of the roughly 4000 “broad field” outputs already produced by the general analysis (that is, the Latin word → shift → English word combinations) and automatically verify which of these are actually compatible with the “single page” constraint.

To explain more clearly: for each candidate word, we could check whether every letter can be assigned to a different page of the magazine, using only the actual available correspondences.

For example, if a word such as “DEMONI” emerges, the test would no longer be interpretative, but purely combinatorial:

does there exist, or not, a configuration in which:
D comes from one page,
E from another,
M from another again,
etc., without reusing the same page?

At that point, the result would become much more interesting statistically, because it would allow us to progressively reduce the number of candidates without introducing manually chosen criteria.

In practice:

• how many of the ~4000 outputs already satisfy the Latin → English constraint;

• how many also satisfy the single-page constraint;

• how many require at most one swap;

• how many satisfy all conditions simultaneously?

I think this approach would be more robust, because it would transform the problem into a well-defined combinatorial test, avoiding at least in part the risk of selecting patterns a posteriori.

I would add one further consideration, however: since we have a large number of magazine pages available, it is possible that many combinations will still turn out to be compatible with the system. Precisely for this reason, I think the interesting point would not simply be whether alternatives exist, but rather how much the number of candidates progressively decreases as the different constraints are applied.