It depends entirely on the card's value ladder, and you can know before you submit.
Look at three numbers for the exact card: the raw near-mint price, the PSA 9 price, and the PSA 10 price. Then run the 9 against your all-in cost.
If the 9 clears raw plus fee, a near-miss still comes out ahead and the card is a comfortable submission. If the 9 roughly breaks even after fees, the whole case rests on gemming, so only send it when the centering and surface are clean. If the 9 sells for less than raw, anything short of a 10 loses money, and that is where flippers get burned: the fee is a bet on a 10.
A useful sanity check is the break-even gem probability: the share of 10s you would need for the expected sale (p times the 10 price plus 1 minus p times the 9 price) to cover raw plus fee. If that number is higher than what your measured centering supports, skip.
Midpoint's price page shows the raw / 9 / 10 ladder with your estimated grade highlighted and a worth-grading verdict that does exactly this math.
Next step: for the next card you are tempted by, write down raw, 9 and 10 before you look at the card, and only submit if the 9 alone does not sink you.
— Midpoint Bot (automated first answer; collectors, please add to or correct this)