I don't. Or at least try not to.

I used to and I lost money fast selling at a loss only to then watch the stock I no longer had go up a stupid amount a few days later. I'm basically an expert at setting my stop loss just above the next swing low before a move up.

If I'm buying stock I'm buying something I have faith in and I'll just average down.

If I'm gambling whatever my maximum risk is I'll just take that money and buy calls with it instead.

The only exception is if I can't afford 1 contract. But then I set my loss at recent swing low whatever that happens to be. I'll then decide how much I'm willing to bet (lose). Whatever percentage my stop loss is is then the multiple for what I'll invest in the gamble. That way I don't have a set percentage.

Example, I'm willing to bet $100. Just under recent swing low is a 10% stop loss. Therefore that's my stop loss and I'm buying $1000 worth. If the recent swing low happened to be 20% I'm buying $500 worth instead.

Or alternatively I'll just buy $100 worth if it's a massive gamble and I can't afford a contract on it and go zero or hero. But obviously in that scenario calls are much preferred.

If it's for the purpose of algo trading though I'll usually stick with 1:1 because I prefer a higher success rate than more money per successful trade. But even then I'm not a fixed percentage as I use SAR for my stop loss value.