Lafferty vs the ciphers

Retired Vallejo officer Lyndon Lafferty has been getting a lot of attention for his book in which he claims to know who the Zodiac killer was. The numerous criticisms of his claims aren’t mentioned by the various news stories, but you can find them here, here, here, here, here, and here.

Lafferty claims to have solved the unsolved Zodiac ciphers, but fails to provide details on his solutions. Like many of his other claims, he expects you to take his word for it. To make matters worse, instead of sharing the solutions outright, he says he intends to profit from them:

The detailed explanation and analysis, including the unbelieveable and undeniable pattern, will be offered for sale at the first opportunity.

Would you pay him?

Well, here’s a taste of his code work: I recently received a copy of his attempted solution to the unsolved 13-character “My name is” cipher. Lafferty appears to take as many liberties in his code work as he does in his book. Let’s look at how he shoehorns his suspect, William Joseph Grant, into the cipher.

Here’s the 13-character “My name is” cipher text:

First, he replaces the crosshairs symbol with the letter Z. Why? Because Zodiac frequently signed his correspondences with that symbol. (Examples: 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18)

Next, he treats the symbols as shift operators. The shift is performed on the symbols immediately preceding them: (Z), K, and M.

Here’s how the shift works:

So, Z K and M are shifted to G R and T.

 

The remaining symbols that look like normal letters are treated like normal letters:

Then, he rearranges the text to read:

NAME GRANT

He then hypothesizes that the symbol is the astrological sign for “Aries”, or the “Ram”, and substitutes the symbol with those words. The words “ARE” and “IS” can be found in “Aries”, and “Ram” can be re-arranged to “MAR” (Grant was born in March). I’m going to ignore this step for now, so we can focus on how Lafferty generated a name out of the cipher text.

His decryption process is very convoluted. Does that by itself make it wrong? Not necessarily. But convoluted solutions tend to have many problems. My rule of thumb for verifying solutions is: Can other reasonable solutions be generated from the same process? If so, then the person claiming their solution is correct needs to explain why all the other possible solutions are wrong. Otherwise, their guess is only as good as any other guess.

As you might expect, if we use Lafferty’s steps, we can extract numerous names other than Grant’s.

Since Lafferty allows anagrams in his approach, we must examine all the possible anagrams. If we rearrange the letters that form NAMEGRANT, we can find many other names. Here are some that I found just by looking at common names in U.S. Census data:

Last names: TANNER, REAGAN, REGAN, GERMAN, RAGAN, ARMENTA, GARMAN, RAMAGE, ARENA, RANGE, EAGAN, MANGAN, MARTE, TERAN, ENGRAM, MAGER, AMANN, GARTMAN, MAGNER, MANER, GRATE, GAETA, RETANA, ARTMAN, TENGAN, AMENT, EATMAN, ARANT, MARTEN, MANNA, GERMANN, ARENT, ENGMAN, GARTEN, MANGER, GETMAN, AMANTE, GRANA, GAMET, MANNER, ANGER, AGENA, MATER, GENNA, EAGAR, RENNA, ANGERT, NATERA, RAMAN, MAGAR, GANTER, ARTEGA, MAGNAN, TANGEMAN, RANTA, MARAN, MANTER, AGENT, TANGEN, MENNA, MARTA, ARMAN, TRANG, ANGERMAN, MATERN, MATERA, NATER, NAGER, MANTE, MAGAT, MAGAN, GENTA, GARATE, TRANE, TRANA, MEGAN, MARET, MAREAN, MANNE, MANGEN, MANGAT, GANTNER, ENAMA, AMENTA, TARAN, NEMAN, NEANG, MAGNANT, GARANT, GANER, GANEM, ARMANT, TANNA, NAMAN, MERTA, MERNA, MATAR, GRANAT, GARAN, ARNET, AGNER, TERMAN, TAMER, NARET, MANERA, GEARAN, ARTMANN, TRAME, TANGREN, MEGNA, MATERNA, MAREN, GANNER, ERTMAN, ENMAN, ARMENT, ANTMAN, ANTER, TREGAN, TANEN, NEGRANA, NAMER, MARANTE, MAGERA, GREAM, GERMANA, GEARN, GEANT, GARMEN, GAMER, AMERT, TEMAN, TEGAN, TARMAN, TANGERMAN, MERANA, MEARA, MATEN, MARGAN, MANTEGNA, MAGRANN, GRAEN, GARNET, GAREN, EARMAN, ARMEN

First names: MEGAN, MARTA, MEAGAN, GRETA, TAMRA, TANNER, TAMERA, MARGE, RENATA, TRENA, GARNET, TAMAR, REGAN, MAEGAN, MARNA, MAREN, TANNA, MERNA, RAEANN, MAGAN, MAGEN, MAGARET, RANAE, TRANG, MARGET, REAGAN, GENNA, REANNA, TAREN, RENNA, TEGAN

How do we know that “GRANT” is the one correct selection from so many possibilities? Is it simply because “GRANT” appears with the word “NAME”? If so, why would the letter read: “My name is NAME GRANT“?

But it gets worse, because we also have to consider several simple and reasonable variations of Lafferty’s approach:

• Lafferty shifts the Z, K, and M to the right to produce G, R, and T. But we can also shift to the left, producing S, D, and F.
• Lafferty also assumes the symbol can be replaced with the letter Z. But the same symbol appears in the solved 408-character cipher, where it decoded to the letter D. If we use D to represent , then it can shift to the right to produce K, or to the left to produce W.
• You can also find a similar symbol in the 408-character cipher, where it decoded to the letter R. Or, you can simply replace it with a T since it looks like an upside-down T. Or, we can ignore the symbol, which was sufficient to produce “NAME GRANT” in Lafferty’s original solution.

If you consider all of those variations, then all of these combinations of letters are possible:

• AENGRTNAM
• AENGRTRNAM
• AENGRTTNAM
• AENKRTNAM
• AENKRTRNAM
• AENKRTTNAM
• AENSDFNAM
• AENSDFRNAM
• AENSDFTNAM
• AENWDFNAM
• AENWDFRNAM

Another quick experiment on U.S. Census data reveals that many more names can be found when rearranging those symbols:

Top 100 of 625 last names found: ADAMS, NEWMAN, GARNER, KRAMER, TANNER, TRENT, EASTMAN, MEANS, REAGAN, MADSEN, REGAN, GERMAN, SERNA, ANDERS, WARDEN, SEAMAN, REDMAN, ARNETT, ANDREW, MENARD, GARNETT, WENDT, MARES, GANTT, DANNER, SANDER, ADAME, SNEAD, RAGAN, ANDRES, FRAME, MEARS, MANNS, ARMENTA, STEADMAN, ANDRE, ERDMANN, MADERA, DENMAN, MARRA, TREAT, FEARS, MARKER, GARMAN, RAMAGE, ERDMAN, WRENN, REAMS, MATTA, RAMER, GARTNER, ARENA, MADER, RANGE, MAREK, EAGAN, DEANS, ADAMES, RAGER, ARENAS, MATTERN, MANGAN, MANNERS, MARTE, DEMARS, TERAN, DEARMAN, ANDES, ARMES, MATTE, DEWAR, STEDMAN, MARDEN, WARMAN, ENGRAM, MAEDA, NADER, MARSDEN, MANKE, GARREN, KANTER, SANNER, DATES, DANNA, ANDREAS, ARMAND, MEADS, MAGER, AMANN, WANNER, GARTMAN, MAGNER, FREDA, ARENDS, ARDEN, SEMAN, MANDERS, WARNE, STEAD, MANES

102 first names found: MARGARET, KAREN, SANDRA, ANDREW, ANDREA, WANDA, MEGAN, ANDRE, DEANNA, ANDRES, FREDA, MARTA, TRENT, ARMAND, MARGRET, DEANA, DANTE, MEAGAN, STEFAN, GRETA, GERMAN, DAREN, TAMEKA, ANTWAN, ANDREA, TAMRA, TANNER, TAMERA, DEANN, TERRA, ARDEN, DANNA, ANDREAS, GARRET, ANNETTA, SANTA, MARGE, RENATA, NEDRA, ANDRA, TRENA, DAWNA, GARNET, TAMAR, KENNA, DENNA, KARAN, RANDA, GRETTA, REGAN, MAEGAN, RETTA, ANDREW, KARREN, TWANA, ARNETTA, MARNA, GARNETT, MAREN, TANNA, MANDA, FERNANDA, ANDRE, MERNA, MARGART, KARMEN, RAEANN, MAGAN, KARMA, DREAMA, MAGEN, DAWNE, MAGARET, DANAE, RANAE, TRANG, DREMA, ANNETT, MARGET, TWANNA, MARGERT, WANETA, REAGAN, MARAGRET, KARENA, TANEKA, TWANDA, TARRA, GENNA, SANDA, ANDERA, ADENA, REANNA, RENDA, DWANA, TAREN, TAWNA, RENNA, MARKETTA, TEGAN, SARAN

We haven’t even begun to look at other possibilities, such as other anagrams that involve words and phrases that aren’t names. Or further variations of Lafferty’s shifting scheme, such as performing the shift on the symbols following the symbols, or starting the shift count on the first shifted letter instead of the original letter. Or including complete word substitutions to the list of letters to rearrange (for example, ARIES or RAM for the symbol, or CROSSHAIRS or TARGET or ZODIAC for the symbol, or TAURUS or EIGHT for the symbols).

Lafferty says, without merit, that his solution “appears to be the absolute solution due to the brevity of characters. It must be a scrambled, plain text code.” This claim is a delusion, since as I’ve shown above, even with a brevity of characters, we can produce numerous names using his approach.

Even if you treat the 13-character cipher as a normal substitution cipher, in which every symbol represents a single plain text letter, and where no rearrangements are performed, you can find many solutions that fit. If you permit rearrangements and other operations that are not part of standard substitution ciphers, the number of solutions goes up tremendously. Lafferty goes a step further and combines several kinds of operations into his decryption method: anagramming, Caesar shifting, and whole-word substitution. When you give yourself this many tools, you can build just about anything.