So, say I'm playing Shane in 9-ball and I can either take a shot that I figure to make half the time but would be followed by an easy out or play a weak safety that I can execute almost every time, I would take the shot. Against a player that isn't likely to be able to do anything with my weak safety, I'd play the safety. It all depends on which course of action is most likely to result in me winning the game. Letting Shane back to the table with a look at the object ball would probably result in a loss, but against a weak player, I might still be the favorite to win the rack.
It's not a question of why a shot that would work against Shane wouldn't work against a weaker player, it's a question of whether it's worth taking a risk on a shot that might work against Shane when there's a shot available the is more likely to work against the person you are playing.
I do get what you're saying. I'm thinking now of a practical example.
You have a thin cut you'd consider a 40% shot on the 8. 9 stuck to the foot rail.
Let's assume any missed shot = lose for either player.
The weak safe will leave a full table bank or long thin cut.
Shane is 70% to make those shots. The APA5 is 10%.
So the reasoning is... against shane I fire, because 40% chance at winning beats
safing and then having only a 30% chance of him missing and losing.
Against the APA 5 I safe because a 90% chance to get back beats my 40% chance if I shoot.
The problem is, the 90% is "artificially inflated" because the other guy is an APA5.
So can you steal a win vs. that 5 by exploiting his weakness? OK, I can admit it... yes.
But do you really care? You beat a guy you're supposed to beat anyway.
Does anyone track or care about wins vs. weak players?
Your odds of beating SHANE or any good player are no better.
And even if you win, you probably know it in the back of your head.
So I'd argue that, for the sake of personal improvement, you should learn to take those 40%
shots so you can learn how to win when you're playing someone with 'real odds' and not 'fake odds'.
Besides all that, I think in most pool situations, the relationship between safe-or-shoot
is not skewed like my example. A more realistic set of percentages will show that
the right shot with the best outcome is the same vs. either player, they just scale up or down.
The best option will still be the best option, it will just have a lower success rate vs. the pro.