Why Hi-Lo Lite fairs so poorly?

I only compared a few games on your impressive site, but here is what I found:

6d, Fair pen (3/4), H17 DAS LS, 1-20 spread, $20,000 br

Hi Lo Lite (all indexes): $57.00
Hi Lo (I18/F4)..........: $56.69
Hi Lo (all indexes).....: $63.80 (with 4 errors per hour)
Zen 98 (all indexes)....: $61.98
Halves (all indexes)....: $72.35

1d, Fair pen, S17, 1-3 spread, $20,000 br

Hi Lo Lite (all indexes): $127.51
Hi Lo (I18/F4)..........: $137.43 (!)
Hi Lo (all indexes).....: $161.10 (with 4 errors per hour)
Zen 98(all indexes).....: $161.81
Halves (all indexes)....: $186.90

Questions:

1)?@I`ve always assumed that rounded, composite indices lose little strength over more exact ones.
But your sims suggest otherwise, or do they?

2) Is it the "True Edge" method of bet spreads that causes the weakness?

3) Further down you mentioned that the RPC would be your count of choice. Is it true that can use all the same indices and bet spreads (described in BJA) for Hi-Lo simply by dividing by half decks?

I ask because I compared the gain per true count point and found that Hi Lo was about .52% and the RPC was about .27%

Close, but not an exact match.

Thank-you for any response.

Cheers!
Syph