I don't seem to very good at finding gridding patterns... I never seem to be able to figure out what sectors they are likely to hit next.

My experience with gridding so far is mostly in unlims.

A good gridder will be hitting random sectors on their border; when they enter my sector, they enter from a sector containing their figs, limps and armids.

So if my twx database is perfectly up-to-date, I guess what I could do is generate a list of my sectors bordering on their mined sectors, and run a script that crosses those sectors off the list as they are hit by the enemy. I can tell the script to plock a sector in the list. Not sure exactly what I'd do if they hit my plock, I guess there are various ways to proceed, but the main problem is predicting what sector they will hit.

The problem with generating this sector list is that it's pretty tough to keep the information up to date. I suppose I could also write a script to visit border sectors and holoscan to update the database before running my plocker....

Is it worth writing such a script?

Your shouldnt be predicting what sectors they will hit. You would predict how there hitting them then build a target list off available sectors within your grid that meet those parameters. It sounds like you would benefit from a tourney or two. Or playing some turn games, unlims are bad examples and therefore dont offer much value in gridding/anti gridding techniques.