I've noticed that there are some games that start out quite fun, but don't stay that way.

For example, The Prestige Tree Rewritten:
* The first 3 rows are quite fun. The game plays pretty quickly, there's multiple routes, and even on the occasion that you have to wait, it's never more than a few minutes.
* But then, starting in row 4, things slow down significantly. You end up going a while without any interesting upgrades, just getting more of the two new units, occasionally getting an upgrade or doing a challenge, but most of the time just repetitively clicking a couple things on occasion, with waits sometimes being necessary.
* And the following row is also a pain, particularly the Balance layer.

I've heard that Baldur's Gate 3 is like this; people say that Act 1 is good, but Act 2 and 3 aren't that great. (The fact that Acts 2 and 3 weren't in the EA/InDev version, and hence didn't benefit from player feedback, might have something to do with this.

So, what are your thoughts about games that are like this?

I haven't played BG3, but I think it's fairly common especially for long RPGs to front-load the best content into the first 10-15 hours, as they know that's how much most reviewers (and likely gamers) are going to play. Also, it's hard to maintain a high level of quality, engaging content for a 60+ hour game. There is going to be a tendency to start to rush and cut corners with development, as time/budget constraints start to bite more.

Imo, game reviewers should always state with RPGs how much of the game they have actually played, because of the front-loading tendency.

I think one of the reasons that BG2 was such a classic is that they managed to maintain a level of very high quality throughout most of the game. Although, admittedly, it becomes noticeably more linear in the second half. I also thought D:OS 1 was quite consistently good all the way through.

A couple of examples of games that I love that nevertheless (imo) show signs of rushed development towards the end:

Morrowind: most of the fortresses in the Red Mountain area are very small, compared to many dungeons that you come across earlier in the game.

System Shock 1: I played this last year - awesome game, but the final 2 levels seemed rushed and less interesting, compared to many of the earlier ones.

So ... development rushing happens.

PTR was -- I think -- the first clicker game I discovered.
Yes it was a slog. But I was so fascinated by the beauty of the big numbers I had to go on.
I guess one needs more than one factor to like a game and then keep going. Novelty or first impression alone won't do it. For me, in the case of PTR, it was novelty (new game type) + topic (sci-fi geek stuff) + curiosity (what comes next?) + admiration (of the flawless mathematical mechanics). Would I recommend it to others? Uh... maybe but probably not. :-p

Must be true then.

Here's another one, just for you:

Some people on the internet say that a majority of video games is never played to completion. Number are are varying of course, claiming up to 80-90% of all attempts are never finished. But even if it is merely 50%+: that is still a very good argument to prioritize the first half of a game in development.

I can list a lot of "good arguments" for (what I consider) questionable practices.

With this I'm not saying that BG3 Part 2&3 aren't very good.

Honestly, I'd recommend playing until you can do the first Quirk reset, preferably getting 2 Quirks from that reset. (If you only get 1, you have to reset too much of the game.)

(I'd also consider that early end a speedrunning goal; start the timer at the first Prestige reset, and end as soon as you get 2 Quirks.)

If I were to try to improve the game, I would probably leave the first 3 rows unchanged, but then tweak the rest so there's a bit more breathing room with the numbers, and so that you don't have to do certain repetitive things for as long.

I finished BG3.
Part 2 and 3 were fine, but some part of the very last segment do feel a bit rushed. Not bad, mind me, but it's clear there wasn't the same care as in the rest.
The patches already fixed much of the worst of act 3 (mainly bugs and bad performance), but I fully expect some enhanced edition or such.

On topic... how can anybody not mention Mass Effect? Probably the most famous letdown of the last decade.

I'll give a game that has peaks and valleys, Soldier of Fortune II.

Prague: A promising start, shows off a variety of gameplay you'll be seeing throughout the rest of the game.
Colombia: Mixed bag. A lot of vegetation and level design that is basically a series of "outdoor-idors" that makes it feel kinda artsy, but rather confined for what it should be.
Finca: Starts off with a broken stealth section that doesn't really matter, but is otherwise a fun shoot fest in a nice looking manor.
Cargoliner: Good first half that is fairly open and stealthy, bad second half that is just a bunch of cramped corridor shooting.
Hong Kong: Decent all around. Colorful urban environments that are fairly open and straightforward shootouts. No real complaints except for one dead end where going forward may as well be a secret.
Hospital: Fun shoot fest that ends with you shooting an osprey. Enemy difficulty escalates here and has good pacing.
Kamchatka: Goes on too long for three level types. First is the wide open outdoor levels, then you get into the cramped, same-y labs, and the fun kinda just goes out the window.
Airport: Fun, probably the best level in the game. Well paced, mixes things up, and then ends with shootouts on a cramped airplane.
The Shop: Absolute garbage. Kamchatka was at least polished. This level is a buggy, unfun mess in dark office environments, a forced crawl that's supposed to showcase a firefight between the good guys and the bad guys, ending with a fight against another osprey but one where you will run out of ammo before it dies because it flies away and will not get into line of sight of where the fixed machine gun is. It is a terrible ending to an otherwise pretty good game.

I know plenty of games in this vein!
* Starbound. Gods, what a tragic thing it is. Especially when you consider that in the Alpha/Beta the universe was your oyster.
* Retro City Rampage. Yeah, this one starts out as a "pretty funny" pastiche of GTA, but it falls apart when the mooching cameos start showing up.
* Forager. Lies about the development aside, this game quickly turns into a boring grind very fast.
* Aquaria. It's a beautiful game! Then they introduce combat mechanics.
* Bit Trip Runner. It almost made me a rhythm game believer! Then it turned up from a 4 on the scale of hardness to a 9.
* Epic Pinball. They lead with the best table. Oops.
* House Flipper* It was all going so nice until they introduced painting.
* Locomotion, Chris Sawyer's. You'd think that Transport Tycoon in the Roller Coaster Tycoon engine would be good, then you look at the sad tiny vehicle roster.
* Tales of Maj'Eyal. It seems like a good idea for a roguelike, then you realize how terrible the starting classes are.
* Vangers. I love the concept of this game. I have never escaped Fostral.
* Wuppo. It rides the edge of being interesting, then fails to shake out on intrigue.
* Several Telltale Adventures: You'll quickly find the poor signposting leading you in circles as you try and figure out where the event flags are.

**House Flipper is a technical fault; I still went though all the missions and sold all the houses but why is painting walls such a pain?

There is a logical reason for the beginning of a game being best.

1) The developer has more excitement and begins with all their effort, compared to the deminishing return slog of actual work. As in, the sparkle and magic slowly leaves the creator when fun turns into a never ending work load.

2) Play testers rarely ever reach the end of a game. Thus the "we need X...." only happens in the beginning of a game and thus the end seems empty. Since nobody but the dev team payed a visit. Which also credits any good end to the dev members or truly good testers.

3) End?? As long as the marketing dupes you into thinking the dumpster fire is "good enough" at the start, they have your money before you figure out you were robbed.

As a caveat; one may enjoy an entire game about the beginning if it has enough parts and randomisation to it. A good example would be Cataclysm dark days ahead(in spite of their devs and broken game).
RimWorld and Unreal World are a few other examples.

Life itself is the most engaging game of this as an example!

edit: yet more typos, I type to fast >_<

Doom 3: Was having fun, having fun, ... then "ok this is going on long. And still going, nothing's happened new in a while and no story content... still stuff? Surely I'm nearly the end." I look up a lefel list, I'm only getting half way there. Quit playing.

This sounds a lot like a self-fulfilling prophecy: players do not play their games through completion, so devs rush the second half of their games. The second half of games is of poor quality, so players give up before reaching the end. Rinse and repeat.

Things fall apart, it's scientific.

Tetris