I think these columns should work correctly, although I haven't tested them thoroughly:
cnt_r_probe_def_action_fold =
sum(if[holdem_hand_player_statistics.flg_t_cbet_opp and lookup_actions_t.action = 'X' and holdem_hand_player_detail.amt_r_bet_facing > 0 and lookup_actions_r.action = 'F', 1, 0])
This says that the player had an opportunity to c-bet the turn (meaning that they c-bet the flop) but only checked, faced a bet on the river and folded. If the only river action is to fold then they didn't check, meaning that they are in position.
cnt_r_probe_def_opp =
sum(if[holdem_hand_player_statistics.flg_t_cbet_opp and lookup_actions_t.action = 'X' and holdem_hand_player_detail.amt_r_bet_facing > 0 and NOT holdem_hand_player_statistics.flg_r_check, 1, 0])
Then you can create a new statistic "Fold to R Probe" with Value Expression:
(cnt_r_probe_def_action_fold / cnt_r_probe_def_opp) * 100
See the
Tutorial: Custom Reports and Statistics for more information on building a stat from expressions like these.
The turn version would work the same way but exchanging turn fields (with _t_) for flop (_f_), and river for turn.