In short, a game of meaningful choices and deep, winding narrative threads.
There's little truly groundbreaking in the content of Baldur's Gate 3, but it's structure and delivery are something worth experiencing even for people with no fondness for the CRPG/TRPG genre or Dungeons & Dragons. All of the praise lavished on Fallout New Vegas for open ended storytelling - despite that game's taxing lack of meaningful narrative structure - is actually deserved in the case of Baldur's Gate 3. Paired with a compelling set of impressively diverging core story lines, the result is a game much greater than the sum of its parts.
If there's any fault of design over execution, it is the relatively minor quibble that even on the game's default difficulty, it is easy for players to find their way into truly punishing scenarios.
Make no mistake, Baldur's Gate 3 is a great game, more than worth your time. Insofar as I focus on faults from here on out, it is only because I feel that the rest of the game would fully justify a 5 star rating in their absence, and almost without question of any kind.
Larian seems dead set on mimicking classic Obsidian in other, less compelling ways though. While there's gameplay depth enough for hundreds of hours and multiple playthroughs, likeable characters, and hateable villains in spades the framework tying it all together is rough in places.
Later areas in the game, at least prior to any major patches, suffer from progressive performance degradation during play capable of slowing even an RTX 4090 powered system, suggesting deeper technical problems in need of solving. Much of the game's performance feels like it was released early as Larian attempted to get out of the way of Starfield, despite the studios insistence on the contrary.
Several relatively rare and insignificant, but none the less annoying, visual bugs accompany the performance issues on the "they'll get to it at some point" pile - including meshes that weren't cleaned properly between completion and deployment and missing or mismatching skin textures for cosmetic items.
Most problematic, however, are the dialogue and quest systems - which account for the docked star in this review. While the choices present a massive array of options for players, the actual mechanical underpinnings of the system regularly come apart at the seams. The result is a dialogue experience that is what we would call in gamedev "non-deterministic".
Normally this is reserved for simulations, especially physics, that run at variable rates and therefore might not produce identical results every time they run. It should be obvious why this is undesirable when it's a fitting descriptor for the main storytelling vehicle in a game dedicating so many resources and so much player time to storytelling.
Notifications of conversation events like party member approval changes or skill checks regularly break in such a way that they don't remain visible long enough to read. Dynamic events in conversations sometimes fail to trigger, and at times dialogue - especially towards the end of conversations - cuts off when it clearly should not. A reliance on NPC characters as vehicles for most dialogue trigger zones in the game frequently places an unintended character front and center for dialogue.
Companion dialogue is the most noticeably rough, regularly resetting and playing again multiple times. The game often becomes "confused", for lack of a better word, about events that have clearly been accounted for by the developers - delivering lines that contradict choices or quest outcomes.
Normally I would brush this off as a combination of unavoidable conflict with Starfield and Larian's need to expand in breadth and complexity beyond their prior experience. However, this persisted through to the final dialogue of the game, where multiple scenes seemed to play out with no awareness of what was going on around them - including the appearance of a character who had just died.
The subsequent hard cut to credits with no fanfare or resolution of other significant storylines despite the presence of relevant characters seems to suggest a fundamental breakdown in the dialogue sequence, or more perturbing a lack of implementation of any kind of narrative recap at the close of a long game.
For as much good as there is, and as much as I find the broader state of the dialogue system somewhat understandable, I have to knock the game when a core system compromises elements of the narrative in a narrative-centered game regardless of my sympathy for Larian in regards to the lose/lose choice they had to make in regards to release timing.
The same, however, can't be said for the quest system. If I had to knock Divinity Original Sin 2 for anything, it would be that quests can be easily broken and the broader system as a whole in OS2 is relatively undercooked. Some of this comes with the territory. Allowing large scale sequence breaking and implementing granular player choice is going to introduce conflicts. This is simply unavoidable, no matter how much some know nothing on social media suggests otherwise.
I would be surprised if a team anywhere, of any size, could build a game with the structure and scope of BG3 that doesn't break at some point. The act of developing these systems is simply at odds with the act of ensuring they continue to function correctly at every step.
That said, the bigger problem here is not that BG3 has quest related bugs, it is that these bugs are largely holdovers from Divinity OS2.
Previously tracked quests disappear from the quest log despite still being active. Inaccessible quests sometimes fail to clear from the log. Quest markers regularly persist on the game's map in spite of the objective they mark being completed. Despite featuring so many options Larian had to up the memory allocation for storing choices, there's no way to manage which quests are actively being tracked, regularly leading to scenarios where an array of mostly useless yellow diamonds render the minimap mostly non-functional.
None of these were a deal breaker for me in Divinity OS2, but their persistence in Baldur's Gate 3 is disappointing in light of the massive expansion of Larian's headcount and funding over the course of the game.
While BG3 is technically still indie, it's not fair to treat this game like the rest of the indie dominated CRPG space. Even compared to other larger developers in the space, such as Pathfinder developers Owlcat Games, Larian is massive.
They've officially entered the same space as groups like CDProjekt Red in developing at a scale that is traditionally only possible through large, legacy publishing. As such, it's both head-scratching and disconcerting in equal measure to see such an important part of the game make little to no progression in comparison to their previous work.
None of this should distract from what an achievement Baldur's Gate 3 is, however. Larian has done something special here, if predictably imperfect. I only hope the allure of ever greater scale and complexity does not distract from necessary refinement, whatever their next endeavor entails.
There's little truly groundbreaking in the content of Baldur's Gate 3, but it's structure and delivery are something worth experiencing even for people with no fondness for the CRPG/TRPG genre or Dungeons & Dragons. All of the praise lavished on Fallout New Vegas for open ended storytelling - despite that game's taxing lack of meaningful narrative structure - is actually deserved in the case of Baldur's Gate 3. Paired with a compelling set of impressively diverging core story lines, the result is a game much greater than the sum of its parts.
If there's any fault of design over execution, it is the relatively minor quibble that even on the game's default difficulty, it is easy for players to find their way into truly punishing scenarios.
Make no mistake, Baldur's Gate 3 is a great game, more than worth your time. Insofar as I focus on faults from here on out, it is only because I feel that the rest of the game would fully justify a 5 star rating in their absence, and almost without question of any kind.
Larian seems dead set on mimicking classic Obsidian in other, less compelling ways though. While there's gameplay depth enough for hundreds of hours and multiple playthroughs, likeable characters, and hateable villains in spades the framework tying it all together is rough in places.
Later areas in the game, at least prior to any major patches, suffer from progressive performance degradation during play capable of slowing even an RTX 4090 powered system, suggesting deeper technical problems in need of solving. Much of the game's performance feels like it was released early as Larian attempted to get out of the way of Starfield, despite the studios insistence on the contrary.
Several relatively rare and insignificant, but none the less annoying, visual bugs accompany the performance issues on the "they'll get to it at some point" pile - including meshes that weren't cleaned properly between completion and deployment and missing or mismatching skin textures for cosmetic items.
Most problematic, however, are the dialogue and quest systems - which account for the docked star in this review. While the choices present a massive array of options for players, the actual mechanical underpinnings of the system regularly come apart at the seams. The result is a dialogue experience that is what we would call in gamedev "non-deterministic".
Normally this is reserved for simulations, especially physics, that run at variable rates and therefore might not produce identical results every time they run. It should be obvious why this is undesirable when it's a fitting descriptor for the main storytelling vehicle in a game dedicating so many resources and so much player time to storytelling.
Notifications of conversation events like party member approval changes or skill checks regularly break in such a way that they don't remain visible long enough to read. Dynamic events in conversations sometimes fail to trigger, and at times dialogue - especially towards the end of conversations - cuts off when it clearly should not. A reliance on NPC characters as vehicles for most dialogue trigger zones in the game frequently places an unintended character front and center for dialogue.
Companion dialogue is the most noticeably rough, regularly resetting and playing again multiple times. The game often becomes "confused", for lack of a better word, about events that have clearly been accounted for by the developers - delivering lines that contradict choices or quest outcomes.
Normally I would brush this off as a combination of unavoidable conflict with Starfield and Larian's need to expand in breadth and complexity beyond their prior experience. However, this persisted through to the final dialogue of the game, where multiple scenes seemed to play out with no awareness of what was going on around them - including the appearance of a character who had just died.
The subsequent hard cut to credits with no fanfare or resolution of other significant storylines despite the presence of relevant characters seems to suggest a fundamental breakdown in the dialogue sequence, or more perturbing a lack of implementation of any kind of narrative recap at the close of a long game.
For as much good as there is, and as much as I find the broader state of the dialogue system somewhat understandable, I have to knock the game when a core system compromises elements of the narrative in a narrative-centered game regardless of my sympathy for Larian in regards to the lose/lose choice they had to make in regards to release timing.
The same, however, can't be said for the quest system. If I had to knock Divinity Original Sin 2 for anything, it would be that quests can be easily broken and the broader system as a whole in OS2 is relatively undercooked. Some of this comes with the territory. Allowing large scale sequence breaking and implementing granular player choice is going to introduce conflicts. This is simply unavoidable, no matter how much some know nothing on social media suggests otherwise.
I would be surprised if a team anywhere, of any size, could build a game with the structure and scope of BG3 that doesn't break at some point. The act of developing these systems is simply at odds with the act of ensuring they continue to function correctly at every step.
That said, the bigger problem here is not that BG3 has quest related bugs, it is that these bugs are largely holdovers from Divinity OS2.
Previously tracked quests disappear from the quest log despite still being active. Inaccessible quests sometimes fail to clear from the log. Quest markers regularly persist on the game's map in spite of the objective they mark being completed. Despite featuring so many options Larian had to up the memory allocation for storing choices, there's no way to manage which quests are actively being tracked, regularly leading to scenarios where an array of mostly useless yellow diamonds render the minimap mostly non-functional.
None of these were a deal breaker for me in Divinity OS2, but their persistence in Baldur's Gate 3 is disappointing in light of the massive expansion of Larian's headcount and funding over the course of the game.
While BG3 is technically still indie, it's not fair to treat this game like the rest of the indie dominated CRPG space. Even compared to other larger developers in the space, such as Pathfinder developers Owlcat Games, Larian is massive.
They've officially entered the same space as groups like CDProjekt Red in developing at a scale that is traditionally only possible through large, legacy publishing. As such, it's both head-scratching and disconcerting in equal measure to see such an important part of the game make little to no progression in comparison to their previous work.
None of this should distract from what an achievement Baldur's Gate 3 is, however. Larian has done something special here, if predictably imperfect. I only hope the allure of ever greater scale and complexity does not distract from necessary refinement, whatever their next endeavor entails.