Character-Limited Patterns?

[-Job-]

With this small number of repeated sequences, your estimated values should be reasonable. The only problem I see is that of the 55 changed sequences with an 8.5% error rate, there will eventually be more than one change in the same sequence, so one would need a bit higher error rate.

My goal was to determine the smallest error rate which could destroy half of the repeated sequences, so 8.5% and 46% would be a minimum.

Although, i now see that i should also have subtracted the number of affected words in each round. This would lower the error rates i gave to 8.25% and 38%.

Also, in Pliny, there are 24 five-word sequences, each of which accounts for 2 four-word sequences. In these cases, a single error can eliminate two repeated sequences.

In Pliny there are also 10 six-word sequences and 4 seven-word sequences, which means that the minimum error rates of 8.5% and 38% would need to be lowered further.

I agree that any type of systematic mutation of line-initial and line-final words could by itself largely justify the number of repeated sequences of four or more words.