TL;DR: Off-by-one counting, skipped letters, double letters miscounted, and confusion between letter E and digit 5 cause most wrong mappings. Slow counting and tool verification fix the majority.
Off-by-one from A=0 habit
Programmers sometimes count A as 0 internally, producing B=1 … Z=25. Entire message shifts. Re-anchor A=1 on paper before encode sessions.
Double-letter drift
In BOOKKEEPER, middle pairs blur—students increment position twice mentally but write one number once. Say each letter aloud while pointing.
Miscounting mid-alphabet
M/N/O/P/Q cluster causes skips. Keep a visible reference strip until automaticity builds—hide chart only after two perfect quizzes.
Letter vs digit confusion
Encoding the word ONE versus the digit 1 in mixed messages. Establish rule: words encode, standalone numerals may pass through unchanged per assignment spec.
Transposition when writing
Correct mental map 12 15 22 5 becomes 12 15 5 22 on paper. Read numbers backward check catches swaps.
Tool direction reversed
Pasting numbers into encode mode produces garbage letters treated as input. Read UI label before every paste.
Compare suspect word
Run single word through letterstonumbers.com when manual result looks “almost English” but off by one letter everywhere—classic shift error not random typo.
Peer review protocol
Partner B decodes partner A encode without seeing original word first. Mismatch localizes error faster than teacher-only grading.
Error log notebook
Track personal mistake types weekly. Most students see pattern—always mess up J/K/L—target drill saves tournament points.
When wrong map is intentional
Advanced homework uses Atbash or keyword. Confirm puzzle type before assuming miscount—decode direction changes remedy.
Personal error log notebook
Track weekly mistake types—always miscount J/K/L versus systematic off-by-one reveals different fixes.
Partner blind decode review
Partner B decodes without seeing original word—localizes error faster than teacher-only grading.
Tool direction reversed check
Pasting numbers into encode mode yields garbage—read UI label before every paste during fatigue.
Double-letter drift fix
Say each letter aloud while pointing during BOOKKEEPER-style words—middle pairs blur easily.
When wrong map is intentional
Atbash or keyword homework looks like miscount—confirm puzzle type before regrading entire paper.
Double-O slip while encoding BOOKLET
BOOKLET contains a double O and a trailing cluster that hides it. B is 2, O is 15, O is 15, K is 11, L is 12, E is 5, T is 20. The correct row is 2 15 15 11 12 5 20. The common slip writes one 15. The decoded result is BOKLET, which looks enough like the target that a fast checker marks it right. Read the row aloud as “two, fifteen, fifteen” and require the repeated word “fifteen” to be spoken twice. A single “fifteen” is the error, even if the rest of the row is perfect.
Circle the second O on the printed word before encoding, not after. The circle is a process aid. Students who refuse the circle because they “know doubles” are the students who drop them. Let them skip the circle only after two clean BOOKLET encodes on different days. One clean encode is luck. Two is a pattern.
The digit 5 written where the letter E belonged
A different failure shows up when the plain word contains an E and the student is also writing numbers. In BOOKLET the E is 5, so the digit 5 is correct in the cipher row. The failure is the opposite direction: a student decoding 5 who writes the digit 5 on the answer line instead of the letter E. The word becomes BOOKL5T. It looks like a typo and it is actually a category error. The fix is a two-column header printed on the page: “letter column” and “number column.” Digits are illegal in the letter column.
Demonstrate with a word that makes the confusion obvious, such as BEE, which encodes as 2 5 5. The decode must be the letters B, E, E. If a student writes B55, point at the header, not at their intelligence. Ask them to rewrite only the letter column. Most of them fix it immediately once the columns have names.
Notebook row that records the letter, not the mood
An error log works when the columns are letter, wrote, should have written, and one cause from a short list: double, off-by-one, digit-for-letter. BOOKLET’s dropped O is cause “double.” BEE’s B55 is cause “digit-for-letter.” A log that says “I was rushed” cannot be practiced against. Mood is real and it is not a category you can drill next Tuesday.
Review the log once a week and pick the cause that appears most often, not the word that was most embarrassing. If “double” appears four times, the next drill is FEE, SEEN, and BOOKLET again. If “off-by-one” dominates, the next drill is a slow count of M through R. The log chooses the drill. The student does not pick their favorite word and call it practice.
FEE as the repair when BOOKLET loses an O
FEE is the short double you use when BOOKLET’s second O disappears. F is 6, E is 5, E is 5. The row is 6 5 5. There is nowhere for a missing letter to hide, because the word is three marks long and two of them match. Decode it immediately. If the student writes FE, they dropped the double in a word that is mostly the double. Stay on FEE until 6 5 5 decodes to three letters on two different days.
Only then return to BOOKLET. Do not “mix in” SEEN on the same day as the return. SEEN is 19 5 5 14, another double-E word, and it can be next week’s confirmation. This week’s confirmation is one clean BOOKLET with both 15s spoken aloud. A log entry that says “double, repaired via FEE” is the sentence you want, not a stack of new words.
