I kinda wonder at the amount of "artificial boost" guys using octave have. Would +100% be out of the question? I mean if you set it up right under octave, aren't you skipping the entering of 40 spaces etc?
My understanding until reading your message was that the Octave entries had been entered with all letters AND spaces. If space were not entered, this would be a kind of word completion, which is clearly disallowed by the contest rules for normal entries:
«Obviously, the use of Graffiti shortcuts is not allowed and neither is the use of word or phrase completion capabilities such as the ones offered by Fitaly for the Pocket PC, WordComplete, TextPlus, Octave, or the one that is built-in Windows CE.»
Not entering spaces would give a clear advantage as it would save 40 keystrokes. Hannes will ask confirmation and if you are right, these entries might be more appropriate in the "word completion" category.
Think REAL WORLD!
On the point you are making on whether this is "real world," I tend to share your view.
For a real world text, most of the time the word you want is not the one that can be selected by lifting the pen. And you are also slowed down by having to watch constantly the evolving interpretation. This seems to parallel my own experimentation with T9 which uses a similar dictionary-based mapping of 9 keys on 26 letters.
Of course, when repeating the same text over and over, this issue disappears because the software adaptation increases the frequency of having the right word first. Also, this knowledge removes the need to be watching all the time at the current interpretation.