Reading the probability disclosure properly changes what you expect. Where intuition goes wrong.
It is tempting to think a 1% chance tried 100 times yields one hit. It does not. Each attempt is 1% again, regardless of what came before. After 100 tries, roughly a 37% chance of no hit at all remains. The sense that it is 'due' after many attempts comes from expectation, not probability. Equally, a lucky streak does not lower the odds of the next attempt. Every pull starts again from the same conditions as the first.
A mechanism guaranteeing the item after a set number of attempts is commonly called a pity system. Where it exists, your worst-case cost becomes calculable. Reading that number before the advertised rate is more practical. No pity system means no ceiling, which is an entirely different proposition.
A calculation saying it takes so many attempts on average describes the average across many people. An individual can run far worse than average, and not rarely. An average of fifty does not mean you will get it within fifty. That is why it is more realistic to plan around the pity limit, or the maximum you can afford, rather than the average.
Deciding when to stop after starting means money already spent distorts the judgement. 'I have come this far, just a little more' is precisely the sunk cost fallacy. Fixing an amount and a number of attempts beforehand, and putting a limit on the payment method, is what holds. A monthly limit is easiest to manage, and pairing it with a rule that you stop paying for the rest of the month once you reach it keeps you steadier.
With a payment method saved to the account, a purchase is a few clicks away. On a shared family device especially, turn on purchase restrictions for minors and require a password. Refund policies differ by game and platform, and loot box purchases are often difficult to refund. Turning on payment notifications by text or email lets you spot unexpected charges quickly.
If a gacha has a pity limit, the worst-case cost is a simple multiplication: the number of pulls to the limit times the price per pull. That price has to be the real unit cost after bundle discounts and bonuses. Subtract the free currency you have saved and any pulls already counted, and you have the maximum you could still need to spend. If you can live with that number, go ahead; if it feels heavy, do not start. That is the simplest decision rule there is.
The same stated percentage can mean different things. Whether it is the chance of any top-tier item or of one specific item makes a big difference to your real chance of getting what you want. If a top-tier result can still be the wrong item, the odds of the one you want are far lower than the headline number. And if the rate rises after a certain number of pulls, the base rate alone cannot tell you the average number of pulls, so read the detailed explanation in the notice as well.
Spending on loot boxes creeps up gradually and is hard to notice from the inside. Adding up a month of purchase history shows the gap between how much it felt like and how much it was. Repeatedly going over your own limit, paying more to win back what you lost, or hiding purchases from people around you are signs that control has slipped. At that point remove the payment method from the account, and if it is hard to manage alone, ask a specialist service such as a gambling help line for support.