The ending is ... okay. Like the rest of Shadow of War's story, it somehow manages to simultaneously feel cool and bizarre. The final act is pretty impressive, but more for its visual flair and its new gameplay than for any narrative payoff.

When I say "final act," I mean Act III: Shadow and Flames. This linear capstone on Act II's open-world orc-slaughter represents the cinematic culmination of Talion's struggle against Sauron.

... but then, the game keeps going. Non-spoiler: no, Talion doesn't defeat Sauron in some alternate-universe retelling of Tolkien's stories. Instead, the Epilogue: Shadow Wars shows Talion continuing his sisyphean toil against the uruk legions. After conquering each of the game world's five fortresses, now, they must be defended!

I did the first defense stage, with attackers somewhat below my level, and it was pretty easy. Then the second, with attackers near my level, and it was kind of hard. I saw the third stage had some attackers above my level. And then I looked up how many stages there are. Ten?

Evidently, Shadow Wars requires a hell of a lot more grinding through orcs, leveling-up, captain-promoting, and warchief-buffing, with absolutely no narrative integration until the very end.

By electing not to continue this tiring routine, I'm departing Shadow of War with a high opinion of it. Building my army and conquering Mordor was way fun. After 40 hours of it, though, I've got no interest in the epilogue -- nor in its awkward-sounding DLCs.

(Blade of Galadriel strips out the Nemesis system? And Desolation of Mordor removes respawning? Nah, thanks.)

Better than: Middle-earth: Shadow of Mordor, Agents of Mayhem
Not as good as: The Legend of Zelda: Breath of the Wild, The Witcher 3: Wild Hunt
Boy, if they try to sequelize this again: Monolith will really have their lore-work cut out for them.

Progress: 95%, level 48

Middle-earth: Shadow of War is a lot like its predecessor -- like, a lot like it.

• Shadow of Mordor's story was well-acted and cinematically cool, but the lore and setup felt dumb. That's still true in Shadow of War. (A "New Ring?" Yeah, okay. But Talion still looks like he could be in a Peter Jackson movie.)
• Shadow of Mordor's mechanics felt very lifted from other games, like Assassin's Creed-style sneaking and Batman Arkham-style brawling. Shadow of War layers on some new mechanics, like Dark Souls-style online integration and ... some more Assassin's Creed-style meaningless collectibles.
• Shadow of Mordor didn't show its whole hand until several hours in, when Talion learned how to dominate and brainwash uruks. Somehow, he forgot how to do this, and Shadow of War also keeps that card hidden up its sleeve for the first several hours.
• I was really impressed by the Nemesis System back in 2014, and I figured the next several years of open-world games would pick it up. As of 2018, this is still true: it's an awesome system and I can't believe more games aren't doing it.
• As repetitive as it can seem, mowing down hordes of orcs is still really fun. Using combos to charge meters to activate special abilities to clear the area, only for more enemies to stream in afterward -- it makes Talion feel like a damn super-hero. It's empowering and challenging at the same time.

Shadow of War has basically the same faults as the previous game, but also the same strengths. And while many of its new features aren't that remarkable - the Shelob memories feel especially like trivial filler - it does bring a grander scale. Simply put, there are five regions (compared to Shadow of Mordor's two), and each one has its own orcish org-chart and its own fortress to siege and conquer.

Taking the first fortress was easy, but by the second, it became clear that hunting down bodyguards and planting spies among the warchiefs are necessary tactics. Middle-earth spycraft!

I'm not sure how well it'll all turn out, considering that Frodo and Sam never saw the Bright Lord on their way through Mordor. But at least for now, I really feel like I'm delivering on Celebrimbor's goal of building an army to rival Sauron's.

Shadow of War is a good counter-example to my recent experience with Agents of Mayhem. The latter had mechanically fun gameplay, but a dumb and unengaging plot, and I was unsatisfied with its sense of progression; I didn't feel like I was really making a difference in Seoul.

Shadow of War also has mechanically fun gameplay, and while its plot is still dumb and unengaging -- the sense of progression is very satisfying. As I slaughter uruks by the hundreds, I know that their corpses are building my way toward a better Mordor.

Progress: Just took my second fortress.

I spent a week with Agents of Mayhem, and I ... really wanted to like it. Its fundamental mechanics are fun enough, and it has a bunch of cool ideas for expanding Volition's typical comfort zone. But all of that ambition is brought low by major missteps -- often, confusing missteps, given what Saints Row IV achieved.

AoM's overall theme is the biggest disappointment, to me. It's set up like a Saturday morning cartoon, particularly G.I. Joe, with a collection of Mayhem's "good guys" teaming up against the army of Legion's "bad guys." It even has some pithy loading-screen tips to parody the old "Knowing is half the battle" PSAs. But the parody is extremely thin: the game's story plays out more like a grown-up retelling of a cartoon. Characters may drop an F-bomb every now and then, but it's otherwise just as melodramatic and predictable as a children's show.

It baffles me why AoM has completely discarded the over-the-top silliness of the Saints Row games. That same irreverence would have worked perfectly here, mocking how the Agents are so geographically distributed, or how the villain always escapes at the last second; but instead of mockery, dildo bats, or Mayor Burt Reynolds, it just has flat characters in a bland setting. And villains that actually escape at the last second.

A handful of the character backstories are funny on paper -- like Hollywood, the actor who needed protection after fumbling with a real gun on set; or Daisy, the roller-derby girl whose intro mission is being blackout drunk. But even in most of these cases, AoM's story is delivered in a no-nonsense tone, and its humor is an all-too-brief respite from the banality of the mission.

So the theme is a bust. At least there's a solid game underneath it, right? Well... sometimes.

The running-and-gunning mechanics descend directly from Saints Row and Red Faction. Third-person aiming has gotten better, but it's still not perfect. Jumping works ... well-enough. Driving is serviceable, but still not as sophisticated as, say, GTA 4.

What makes AoM's action-gameplay interesting is its set of damage strengths and weaknesses, along with diverse buffs and debuffs. As the challenge level ramps up, you'll need to know how to use the right characters, abilities, and tactics against various enemies and situations. These fundamentals work really well, and learning them is a few hours of good fun.

Unfortunately, these mechanics are let down by the lackluster content surrounding them. Said content comes in three flavors:

• Dungeons, which are thoroughly, overwhelmingly bland. They're always full of empty corridors and samey battle-chambers. They're so repetitive that I assume they were randomly generated from a small selection of rooms and hallways.
• The city of Seoul, which - while it's an "open" world - isn't fun to explore, due to being crowded by tall buildings. (That'd be fine if you had Saints Row IV superpowers, or Crackdown super-jumps! but you don't, so buildings just get in the way.)
• And some isolated events and hot-spots in the open city, which are as copy-and-paste as the dungeons.

I felt like the environment design wasn't engaging enough to support the amount of time I spent in it, especially in dungeons. AoM employs amateurish techniques of stretching its content further: like filling dungeons with room after room of wave-based encounters, or missions that ask you to drive all the way across the city, then back again. Or the Ark headquarters, which frequently interrupts the flow of the game.

Want to change your three-agent loadout? Need to jump to the Ark, then warp back to the city -- with a loading screen on each. Want to build and equip a tech upgrade you just found? Need to jump to the Ark, then warp back to the city. Want to send an agent on a new Global Conflict mission? Need to jump to the Ark...

That "Global Conflict" mode, by the way, is the worst attempt I've seen yet at an asynchronous time-management minigame. You send some agents off to do a timed mission, after which they bring back some reward and unlock more missions. Right? Assassin's Creed has been doing this since Brotherhood - in 2010 - and AoM's implementation is worse, in difficult-to-understand ways.

• The Global Conflict world map is split into multiple regions. To start, the East Asia region is unlocked; all the rest are locked. To unlock a new region ... you need to do all the missions in your current region. Hence: you can only have one region unlocked at a time. Why even have a world map?
• Agents don't gain experience points when they're sent on these missions. They're just unable to be deployed. (And remember, you need to go back to the Ark after the mission to get them back!)
• And once you finally finish all the regions, you can storm Legion headquarters in Moscow ... which is yet another bland and samey dungeon, with no narrative significance whatsoever. After finishing the HQ, the Global Conflict map resets, "allowing" you to start it all over again.

Even the checklist-ey nature of AoM's open world - a quality I usually enjoy for its own sake - is sabotaged, seemingly, to make the game more "endless." Several locations that you can conquer or claim in Seoul are reclaimed by Legion, as a consequence of continuing to fight them in the city.

Between the insipid story, dull environments, and continuously-reset objectives, Agents of Mayhem fails to establish a meaningful goal to play towards. Other than perhaps unlocking all the agents. Having done that, my motivation to continue has dried up.

Which is a shame, because the core action mechanics of the game are really not bad. They just aren't enough to keep me interested in the rest of AoM's campaign.

I hope that Volition's next game gets the creative budget (and time) it needs to build on AoM's half-baked ideas.

Better than: arguably, the first Saints Row (X360).
Not as good as: Saints Row 2, The Third, IV, or even Gat Out of Hell.
I failed to care about most of the agents: but I might spring for a spinoff game about Daisy.

Progress: 73% -- Campaign: 60%, Agency: 94%, Assets: 65%

Momodora: Reverie Under the Moonlight does some things well. The combat mechanics are easy to learn, and have a bit of satisfying depth; upgrades come at a good pace and feel meaningful; regions have distinctive visual and audio themes. And the map is a good-enough example of non-linear Metroidvania layout, given how short the game is.

But - despite how short it is - Reverie fails to make a meaningful impression. Its mechanics aren't special, its art style is unremarkable, and its story is threadbare, supported by a handful of impersonal lines and a small cast of incredibly flat characters. Even with NPCs, this game feels deader and emptier than Metroid or Super Metroid.

Without interesting gameplay or narrative stakes, the map itself was the only motivation I had to keep playing. And that isn't a great motivation.

Only one anecdote stands out in my memory, and it isn't a good one. I got to the final boss, killed it, and then ... died, cinematically?, clearly having received a "bad ending." But as I could only muster the slimmest margin of interest, I looked up how to get the "good ending" online, which was to backtrack for a groan-worthy hidden item, then re-fight the boss. It doesn't feel reasonable to me that a needle in a haystack is the only way out of a disappointing ending.

(Granted, even the good ending is pretty bland and unfulfilling.)

Momodora: Reverie Under the Moonlight is competent enough in its fundamentals, but just doesn't have anything noteworthy to say for itself. I don't regret the three-and-a-half hours I spent with it, but have no interest in hunting down its secrets -- nor in trying the rest of the Momodora series.

Better than: Shantae: Risky's Revenge - Director's Cut, if only just.
Not as good as: Guacamelee! Super Turbo Championship Edition, which wasn't really my favorite, either.
Hard to say if better or worse than: Ori and the Blind Forest: Definitive Edition; at least that was memorable, albeit frustrating as shit.

Progress: Got the True Ending, 98% map completion.

At its best, CrossCells feels like an inductive-reasoning "expansion pack" on puzzles like sudoku and picross. Like the other 'Cells games, it incorporates multiple mathematical mechanics, and plays their hints off one another. And as I worked through its later puzzles, I felt very satisfied with their logical tightness.

But while CrossCells is pretty satisfying when it "works," it just as often "doesn't." Whether because I hadn't learned its rules adequately, or because the rules weren't applied rigorously enough - I couldn't say, in retrospect - many levels felt like I had to guess to move forward. Only after following that guess for several steps would I know if I was on the right track.

CrossCells would really have benefited from a "try it out" system, like some picross games have, to follow a hypothesis for a while and then undo it if necessary. But, let alone that -- CrossCells doesn't even have a "reset" function. Its interface is as minimally-functional as can be.

(Made a bunch of wrong moves? To reset the puzzle, you need to exit to the menu and click the puzzle again. And the annoying water-ripple animations draw this process out to several seconds. Ugh.)

CrossCells shows mechanical promise, but just isn't quite friendly and sophisticated enough (both in UI and in its difficulty-ramp) to be very noteworthy. And it's pretty short, to boot.

Better than: Hexcells, SquareCells
Not as good as: Hexcells Plus, Hexcells Infinite
... but for a sale price of under $1: it's hard to go wrong.

Progress: Finished all 50 puzzles.

I play different puzzle games for different reasons. Some, like a picross game, are relaxing ways to unwind. Some, like Portal 2 or Antichamber, are fun to learn and explore -- as a journey of understanding the designers' ideas. And some, like SpaceChem or TIS-100, humble me because they make me feel like a complete idiot.

Cypher is in that last category. It really makes me feel stupid. And I love it.

Surprising, at least to me, considering its developer is known for small and relatively-simple puzzlers like Hexcells. Cypher is a huge step up in terms of complexity: after spending a few hours with it, I legitimately believe that three-letter-agency cryptographers might respect these puzzles. Forget about my praise for Nonogram's in-game hints -- Cypher is praiseworthy because you need a pencil and paper to work these guys out.

And it's a step up in production quality, too. Strangely it still uses the default Unity taskbar icon... but the minimalist 3D world, and soft piano soundtrack, form a fantastically soothing library-like aesthetic around these hard-as-shit crypto puzzles.

Zach Barth and Jonathan Blow have a new peer in my esteem of puzzle-creators.

Progress: 039%

Nonogram - The Greatest Painter has an uninspired title, is inexpensive, has "Mixed" reviews on Steam, and is from a publisher that seems to churn out puzzle games. So, if it wasn't for the fact that I am a total sucker for picross, I never would have even bothered with it.

In fact, Nonogram has features that meet or exceed any other picross game I've played. And I've played a bunch! (As for the reviews, it seems like the game had issues at launch that have since been fixed.)

The main mode is a cool-looking "Gallery" of 126 puzzles: starting from the center image, completing each puzzle unlocks the adjacent images. It almost feels like a choose-your-own-adventure campaign, albeit a short-lived one. These puzzles go up to 25x25, which is pretty respectable, although a lot of them are only 5x5 or 10x10.

In addition to that mode, there's a menu of "Classic" puzzles, the first of which is 50x35. I am really excited to dig deeper into that mode. And there's a Speed Drawing mode! like the Time Trials in Pepper's Puzzles, but more of them. Plus a Steam Workshop mode that I probably won't bother with, but hey, well done guys.

Some picross games must limit their puzzle size based on screen layout -- very few offer any zoom at all, and if they do, it's only to pre-defined sizes. Nonogram has an adjustable zoom, with scrollbars! (You know, like any other PC application would have.)

Not only does it gray-out its numeric hints as you complete segments (take that, Picross Fairytale) but it even highlights hints that are relevant. If a row is ready for a new mark or cross, the numbers will be highlighted, suggesting a place for your next move. It doesn't make the game feel simpler or easier; just saves the time that I would otherwise spend tediously scanning each unfilled row and column.

And there's even a little wit in here, in the puzzle titles: some oblique pop-culture references, even some puns. Whereas Pepper's Puzzles sometimes felt like they were just ripping off fan-art, Nonogram's titles make its references feel more like knowing homages.

... granted, the final images take some extreme artistic license, compared to the actual puzzles. But I've got no problem with that.

Throw in a peaceful ambient soundtrack, and there's almost nothing wrong with Nonogram. (The hint highlighting is occasionally inaccurate, thankfully on the side of under-helping rather than wrongly-hinting.)

For how humble it may appear, Nonogram is having no problem besting my previous favorite picross games.

Progress: 126/126 Gallery, 1/50 Classic, 29/50 Speed

Seven years later, Portal 2's campaign still holds up fairly well.

It did surprise me that it took so long before I got to Cave Johnson. And that the Cave Johnson parts were over so quickly. In my memory, this - wandering through the abandoned archives of Aperture - was the bulk of the game.

But in replaying it, the archives were actually a pretty minor segment, compared to all the stuff before (routine Portaling) and after (Wheatley's insane shit's-on-fire puzzles). That other stuff is still fun, both in puzzle design and in humorous writing and voice acting; Cave Johnson is just so much of a show-stealer, that the rest of the game isn't as memorable.

Compared to the thrilling action-adventure jaunts of today, Portal 2 is a little on the short side (5-6 hours), and takes a little while to warm up. But it's still perfectly enjoyable and worthwhile.

Better than: Portal
Not as good as: The Witness's first-person puzzles, and Borderlands 2's humor, both somewhat debatably.
Last time, I wrote: "I look forward to even more from the Portal franchise." I still do, I guess, though we may not live to see it.

I don't know if I'm getting this title right. Steam calls it "Picross Fairytale - nonogram: Red Riding Hood Secret" which sounds like a poor translation from ... something. And that fits with the general production value of the game.

That is, like it was ported from a mobile game, and not very well.

This is the only picross game I've played, so far, in which dragging the cursor over an already-filled square can overwrite that square. Resulting in already-checked or already-crossed squares becoming easily mislabeled. Considering picross games have been fairly widespread for some decades now, this feels like a rookie mistake.

And otherwise, the game has nothing special going for it. The plot is threadbare, a minimalist reference to stories of princes and princesses; the level progression is straightforward and inflexible, pushing you from one level to the next with no surprises; and the interface is about as dumb as possible, with no graying-out of numeric hints when they're satisfied. There is an awkward indicator when a row or column is "right," but since it doesn't take any other rows or columns into account, it's about as helpful as looking at the squares yourself would be.

And you can pay real money for some kind of currency. I don't know what it does. I can't imagine why, in a puzzle game, paying money to ... skip puzzles, I guess? ... would be even remotely okay.

Even considering that it's free, this isn't worth it. You'd do better to shell out a few bucks for a picross game that isn't as outright shitty.

Better than: Pokémon Picross
Not as good as: Pepper's Puzzles
Slightly worse than: Regency Solitaire

Progress: Got to puzzle 8.

Shaq Fu: A Legend Reborn ends up feeling most like a parody game. And it isn't the worst parody-game I've played.

Its mechanics are simple, but basically functional -- at times, even "fun." And what I appreciated the most in the first stage was the game's off-the-wall humor.

But Shaq Fu 2's gameplay gets stale real, real fast. Levels have no meaningful interactivity, there's very little variety in enemy behaviors (most of them are total pushovers), and combat fails to evolve over the course of the game. Actually, of the two (!) power-ups Shaq can find, the second one manages to be even less interesting than mashing the punch button.

Most of the time, Shaq Fu 2 doesn't know what to do except throw wave, after wave, after wave, after wave, after wave, after wave of enemies at you, for minutes at a time.

... and sadly, the game's sense of humor also peters out pretty quickly. Its references to celebrities felt just a little too "safe," like the team was afraid of drawing too much attention to them. Most of the jokes felt low-effort. (I would have expected more of the "Shaq's Chinese?" running gag, which was weirdly under-utilized.)

Shaq Fu 2's troubled development story - which the game attempts to lampoon, to awkward result - may really have been a blessing in disguise. As boring as it was to play through the game, it was still, thankfully, only three hours long.

So, file Shaq Fu: A Legend Reborn under "mocks game design mistakes that it has itself made." And like I said, not the worst of those I've played.

Better than: Eat Lead: The Return of Matt Hazard
Not as good as: Marlow Briggs and the Mask of Death
Probably also better than: Shaq-Fu, but, c'mon.

Progress: Finished on Normal.