Code:
1. 20/13 Eq.: +0,096
0,522 0,157 0,008 - 0,478 0,137 0,004 CL +0,068 CF +0,096
2. 9/7 8/3* Eq.: -0,040 ( -0,136)
0,489 0,161 0,007 - 0,511 0,169 0,006 CL -0,029 CF -0,040
3. 9/4 6/4 Eq.: -0,046 ( -0,142)
0,484 0,170 0,007 - 0,516 0,167 0,004 CL -0,026 CF -0,046
4. 21/16 6/4 Eq.: -0,624 (-0,720)
0,408 0,133 0,009 - 0,592 0,359 0,047 CL -0,449 CF -0,624
Truncated cubeful rollout (depth 10) with var.redn.
147 games, Mersenne Twister dice gen. with seed 765213078 and quasi-random dice
Stop when std.errs. are small enough: ratio 0,1 (min. 144 games)
Play: world class 2-ply cubeful prune [world class]
keep the first 0 0-ply moves and up to 8 more moves within equity 0,16
Skip pruning for 1-ply moves.
Cube: 2-ply cubeful prune [world class]
The thing is not that slotting and leaving blots is bad. There is no connectivity anymore, and black would like to use that pile on his 6-point to hit loose. In other circumstances he would like to start a new life at white's side of the board by recycling, and most of us are wired to do that. Even that 21/16 6/4 can turn out very well, and it's evaluation shows how bad the bot is in assessing chaotic positions. Though there is a realistic chance that white will hit twice or hits and covers, white's homeboard is so deplorable that black can hardly do anything wrong.
However, black has got a better strategy here. The white checker can be captivated behind black's block, so this should be left intact. Playing 20/13 has several plusses. It's not so bad if white hits, because then black can recycle, and if white hits on the 13-point he loses his control point. In case white doesn't hit, if black is lucky to throw a 6 he can make the 6-prime, and otherwise can attack the checker.
NOTE: My analysis is arrived at after bot experimentation.