I was pondering the old Korean game four field kono. This is an awful game with an intriguing type of capture. Actually I forgot and misinterpreted the capture; I thought you jump over an opponent to land on a second opponent. So what if you combined (my interpretation of) the four field kono capture for some pieces plus other pieces that do capture by jumping over, like checkers? You make holes in formations, then have the pawns take advantage of that.
This quickly switched to chess knights and a bunch of pawns on each of the two teams. I never did try four field kono capture. The pawns have unique rules. The board is smaller.
I tried several different setups and board dimensions, but just wound up back with the original idea, which is a 7 x 5 board and a Dablot Prejjesne setup.
The gameplay for Possession is interesting, and I played it many times. However, the ending is dreary, and surprisingly even. The game needed a goal other than attrition. I tried getting pieces to the other side of the board, or planting yourself on the centre cell, or an attrition points system. As of writing, the goal is to take a ball from the centre position. You aren't really taking it somewhere; the game ends when one player can not take possession of the ball and resigns.

The ball was a dime. The hidden dime is awkward and extra work to move around. I changed this - now called a crown - to an invisible piece.

Possession has what I call the A/B issue. Most of the pieces are on A rows, with a minority of the pieces on B rows. In practise it is not a problem.

I've tried many times to try out a new setup, or a new board design, possibly a new approach. Driving me nuts. It looks like there should be a solution to the A/B problem, an even number of rows (and columns), but nothing improves things. We'll stick with plan A.