The major confusion seems to be in what TD_Words (NDW) counts. As the name suggests it counts number of words, not utterances. Bellow is your sample file that I modified to better show what TD_Words counts. The red words are not counted and the blue shows running word count. The output is below the sample data and the TD_Words is 11.
*CHI: it's an apple [% #words 3] .
%mor: pro:per|it~cop|be&3S det:art|a n|apple .
%gra: 1|2|SUBJ 2|0|ROOT 3|4|DET 4|2|PRED 5|2|PUNCT
*CHI: red [% #words 4] .
%mor: adj|red .
%gra: 1|0|INCROOT 2|1|PUNCT
*CHI: two [% #words 5] .
%mor: det:num|two .
%gra: 1|0|INCROOT 2|1|PUNCT
*CHI: oink [% #words 6] .
%mor: on|oink .
%gra: 1|0|INCROOT 2|1|PUNCT
*CHI: xxx .
*CHI: tomorrow we will [% #words 9] .
%mor: adv:tem|tomorrow pro:sub|we mod|will .
%gra: 1|3|LINK 2|3|SUBJ 3|0|INCROOT 4|3|PUNCT
*CHI: okay [% #words 10] .
%mor: co|okay .
%gra: 1|0|INCROOT 2|1|PUNCT
*CHI: &-eh .
*CHI: two [% #words 11] .
%mor: det:num|two .
%gra: 1|0|INCROOT 2|1|PUNCT
Verbs_Utt TD_Words TD_Utts TD_Time_(secs)
0.143 11 9 0
Leonid.