A shrewd Baldur's Gate 3 fan has discovered that the RPG's D&D spell tooltips might have been bugged for months, which could explain why Shadowheart never lands Sacred Flame

Sacred Flame and Firebolt failing to land? Don't worry, you can blame it on Baldur's Gate 3's bugged spell tooltips! Well, not really. I myself rely mostly on percentages, but if you've been using BG3's tooltips to help inform your chances, you might've noticed a visual bug caught by not one but two Redditors. 

"Today I learned that Sacred Flame doesn't suck, the game just lies," writes UncommonEra on the official Baldur's Gate 3 subreddit. The cleric cantrip is infamous for its slim chances of actually hitting, with the attacked party needing to succeed a dexterity saving throw in order to evade the spell. But in Baldur's Gate 3, it appears a UI bug or "coding issue", as UncommonEra suggests, gives inaccurate information in the spell tooltips as you hover over each spell. If you're someone with staunch D&D and tabletop roots, this could be the reason you're struggling to calculate the exact odds – and it could be why you keep expecting Shadowheart to have better chances with Sacred Flame.

"The tooltip states that your Spell Save DC [damage class] is based on YOUR DEX modifier, rather than your spellcasting modifier (usually WIS). The success chance even reflects this," explains UncommonEra. "However, if you check the combat log after casting, it shows that the calculation was in fact made with the right ability modifier." In short: the tooltips are seemingly giving incorrect information, especially regarding damage classes and which stats are being used for rolls against them. It's important to note that the damage itself all adds up – it's just that pesky tooltip throwing us off our calculations.

Another Reddit user shared a screenshot [via Imgur] to demonstrate the bug, showing three different values for the player's spell save DC when, according to their calculations and attack log, the numbers should have been uniform.

The tooltip bug seems to have been around for some time now. Six months ago, Reddit user plus_300 posted a now-archived thread, asking why