Assassin's Creed IV: Black Flag - although this is mainly the intro section being merely okay before the game opens out into a pretty great epic pirate fantasy. A game from the days when UBIsoft were (despite their DRM stance) a good developer, rather than the mess they are today.

I also liked the second half of Outer Worlds slightly more than the first half. I do genuinely think it's the best RPG I've played in years.

Dishonourable mention to Assassin's Creed III. Apart from a twist half way through, the first third of the game was pretty crap.

On the more traditional front of games starting great and then becoming average (I know it's off topic for this one) - Mafia III. It's an amazing opening, a great world and then... It's just a bit dull and repetitive.

Indeed, it is far more common to have games get worse later on, or at least seem to, but on top of an amount of fatigue setting in, like BreOl72 said, and there being a limit to the amount of creativity the devs can typically pour into a game, there's also the fact that they can afford to make the later parts worse. If the early part of a game is poor, many won't get to the possibly better later one, while if the game gets worse towards the end, not so many would be willing to give up after having committed to it to that point.

That said, in terms of mechanics, D&D-based games like those already mentioned (or D&D-like, such as KotOR - just the first though) can do well, because low-level D&D is so uninspired. And it could also apply to TBS games with RPG development of hero units, particularly notable in those where the way to play is an army of heroes, like HoMM4 or AoW2.

And, as I kept saying whenever dtgreene brought it up, I like the inverted difficulty curve. But it needs to open the way to great exploration / story / lore / characters. If you don't have that and the game also becomes a walk in the park, there will be little left to it. But if you have that early pain for later gain style, where early on you struggle to push through, focusing on that, then get good character(s) and the direct challenges of the game become far less, well, challenging, you can focus on the world and story and can really shine... assuming that they actually do, of course...

PS: An example here, Two Worlds (the first one). What that game has going for it is the exploration, and once you get a sufficiently pumped up poison cloud, and maybe some good healing to account for packs of ranged enemies, you can just explore, enjoy the world, since that spell is ludicrously overpowered.

I'm going to add Lego Harry Potter: Years 1-4 to my list. I think part of it is that The Philosopher's Stone isn't the best book in the series, but it gets better as it goes along. The level design in years three and four is just brilliantly fun. Plus, it's highly sympathetic to Rowling's world (and genuinely funny).

Rune has a famously blah start with the dreary caves, and gets noticeably better after a few levels.

Some other examples:
* Dragon Quest 6 and 7. You don't get access to the class system until something like 15-20 hours into the game, but once you do, the gameplay possibilities open up significantly. (Unfortunately, DQ6 has you fight a major boss that's rather difficult right before this, and DQ7 has you go through a rather annoying section first.)
* Final Fantasy 6 gets far better once you get the second airship. At this point, the world opens up, the game stops bombarding you with cutscenes, and you can choose what order you want to do things in (or what you just don't want to do, though you may miss out on party members for doing so). (Unfortunately, the game becomes less fun when you get the Ultima spell, which is far too strong; that spell really should *not* have had the "ignore magic defense" property.)

I think the Legend of Heroes Trails in the Sky would fit this. The first game in any of the trails arcs are doing a ton of world building and setup to make you know about and care about the people and places involved. This can make the first half of the games seem slow and boring. But once that has all been established it moves at a break neck pace and is all action till the end with people and outcomes you'll actually care about.

Ahhh, you beat me to it! xD

Baldur´s Gate has two very different parts and if not for what happens right after leaving Candlekeep, I would say that it also has two very different moods. The first part is a series of light hearted adventures. Which is nice. And the second one, yes, it is epic, but what I like is the very well told story. Which by the way, is way better than the one in Baldur´s Gate 2.

Baldur´s Gate 2 is an orgy of high level combats. Which is nice, but makes it silly. And Athakatla is filled with all kind of monsters: Lichs, mindflyers, vampires, beholders, etc. Everything together under the city. It doesn´t make sense. It would be great if there were more cities instead of having the main city as the big hub of the game.

I never liked the underdark.

By the way, Baldur´s Gate has a better design than Athakatla.

I feel like that's almost every other (W?)RPG out there. The more "epic" they're supposed to become, the more tedious they tend to be, endlessly drawing out the final confrontation by making you wade through armies of same-y opponents, and the effect is the opposite of feeling "epic" because it's just slowing things down with filler content.

Anyway, regarding the topic, off the top of my head I can't think of any game right now that was better in the second half, but there definitely are games that take a while to get going. I'm currently playing Dread Templar and the first episodes were pretty standard fare; locations and opponents get more interesting and unique later on. (But I haven't finished it yet, so I can't tell whether the second half is better or whether it's just the middle part.)

Hard agree. I think Portal 2 especially opens up in the second half, once you hit the basement and they introduce the paints especially.

JRPGs can also have this same issue.

I think the best solution is to increase the enemy variety in any given area, and make different enemy types actually feel different to fight. In a high-attrition game, for example, mixing weak and strong enemies can work well; the player then needs to figure out which enemies are worth spending resources on.

Also, varying enemy types at different parts of the dungeon can also help. Final Fantasy 1's final dungeon, for example, has elemental themed floors, each with a completely different mix of enemies.

There's also the issue that not every RPG needs to be a 60+ hour epic. In fact, any that have irreversible build choices should really be much shorter than that, so the player can more easily try out different builds.

This would help for sure, but it isn't an insta-cure. One game I know I had this problem with was Pathfinder Kingmaker, which had notoriously tough endgame that went on and on and on. It introduced new enemies and a new area and all that, but that didn't stop that ending from being a slog.

I tend to play RPGs and CRPGs especially on their harder modes, but I also tend to turn them down to normal or even easy toward the end when I just want to finish the game and its throwing endless crap at me.

I think I did go back and finish both Pathfinders on hard though, iirc. Let me check my chievos.

Already mentioned so little repeat

Fallout 2 - Vault City onward is when the game becomes great.

KOTOR 2 - Not really second half but more like 2/3rds. after Pegasus station and the first planet the game really opens up, ending is broken tough.

Witcher 1 - for most first half of the game is not that good. I loved it being slavic myself and feeling nostalgic to my home country when playing while leaving overseas.

Half Life 2 - I find the game to be incredible even if now I speak with nostalgia (did beat it again few years ago) but the moment you get gravity gun and onwards the game is epic. The tunnels and hovercraft are great still but are much weaker than second half of the game.

Hello StingingVelvet,

"Name games where you feel the second half is better than the first. Games where the later sections made you excited to keep going, rather than push through to the finish."

Curiously, no one mentioned the pletora of Zachtronics' problem solving (programming) games, yet.
I think, games such as "Spacechem", Infinifactory", "TIS-100", "Opus Magnum" and so on do fit the criteria fairly well.

Although, some of them allow the player from the very beginning to use all availables tools ("TIS-100" for instance), while others open up the tool box step by step (e.g. "Spacechem" and "Infinifactory"), both ramp up the variety, difficulty and complexity of the tasks during their progress!
And then, there is the can of revisiting former levels with the newly gained tools for alternative solutions--not even mentioning the box of pandorra, that is the attempts to optimize your solutions afterwards.


For those who does not consider the Zachtronics and alikes as 'games' but more as chores or work, I have another suggestion:

"Lumo" A lovely homage to the old school "pseudo" isometric platform adventures of the ZxSpectrum and so forth.

Why, I consider the second half or so of "Lumo" better than the beginning? Because it does not only increases the level of complexity and difficulty (Area navigation and obstacles) but it literally opens up the initially linear level structure to a more non-linear one.
The first half of the game is filled with nice and clever references to many older games of the sort. However, the later non-linear sections return gamplaywise more to the original games which "Lumo" is an homage to--"Knight Lore", "Alien 8" and "Head over Heels" were all rather non-linear in their level design.
Furthermore, "Lumo" does mix up the gameplay in the later sections with various (semi-)optional minigames and levels inspired by different genre classics, such as "Pacman", "Marble Madness", "Zaxxon" and others.

In my opinion, "Lumo" is a very underrated gem, especially, if you have played or know some of the referenced classics.

Kind regards,
foxgog