I believe it was to make the 0-9 unit harder to track and calculate the match at the end of the game. It was a way to better "randomize" the match. I think my Rack A Ball does the same thing.

/Mark

I stumbled across this last week, on Bank-A-Ball, while trying to fix the bell for 5 point scoring. Saw the difference between the point scoring and bell ringing on the schematic and with a test lamp. Turns out my issue was too much travel on the clapper - it rested on the bell when the coil was energized.

But... then I checked out a few other machines’ schematics around the BaB timeframe and found a suprising variety of techniques to do, what John Osborne calls, “Match Mixing”. Of the six machines I looked at, there were five different techniques used… from point filtering to injecting additional signals such as a pulse from the 50 point relay and from the game feature circuits. Some machines drive the 0-9 unit with the 10 point relay instead of the 1 point relay...

It's not difficult to predict the match number on games that don't randomize, and have perfectly operating score reels and steppers. At a recent show, I was showing someone how to do it, and most of the time we were able to accurately predict the match number.
We also determined that sometimes it's impossible for the game to match, as the score reel and the number (hidden during game play) spotted by the match stepper, do not intersect. A person among us erroneously stated that the odds of matching are one in ten. That's not at all true. I guess in rare cases one could steal a replay by allowing the last ball to drain at an opportune moment, but I'd rather rely on skill to win a free game.

Predicting matches was a thing...
Match predictor (resized).jpg

Fascinating thread. Thanks to the OP and all those who shared their knowledge.