FLAC: weak encoding looks good

Hello,

Yesterday, i realized that file sizes obtained from the
pre-defined compression settings (-0 to -8) were rather tight.

For the sake of curiosity, i also tried even lower
"rice partition orders": "-l0 -r1" and even "-l0 -r0".
and i was surprised to realize the file obtained was smaller
than the one obtained with pre-defined setting "-0".

I dont know what's behind "-r0" and even if that value
does make sense (was it really the value used) ?!

Also, i wonder what are the effects of (-b): what are the
pros and cons of raising/lowering the value of 'b' ?
Does this mean that more/less samples are interpolated
by some fixed function ?

Phil