7 Billion Humans isn't "just" a sequel to Human Resource Machine; it also expands the fake-programming domain to multi-threading. Your program runs on multiple agents/workers at once!

Unfortunately 7BH doesn't fix the biggest problems I had with its predecessor: dragging-and-dropping instructions is a chore, the instruction set is awkwardly limited, as is its concept of memory or variables... and this game's new complexities exacerbate a voodoo-polymorphism problem that the previous game merely hinted at.

Egotistical professional programmers (like myself) often look down on languages like PHP and JavaScript because their type systems, or lack thereof, inhibit strict definitions of a program's expected behavior -- encouraging unexpected behavior that's difficult to debug, or even to detect. 7 Billion Humans may keep its instructions "simple" by dodging the question of type-safety, but this leads to haphazard consequences like step sometimes not working depending on the state of another worker in the destination, or giveTo throwing your worker into a shredder if it isn't holding anything.

Personally I find this kind of unpredictable mechanic tiring and unsatisfying - in programming or in video games generally - and this eventually dulled my interest in solving the game's ongoing puzzles. I didn't even get far enough to unlock functional synchronization instructions; in most of the puzzles I completed, 7BH's multi-threading concept was more like a proxy for running the same program over multiple data sets.

7BH also doesn't address my most-superficial criticism of HRM, that it doesn't compare your solution's memory- or runtime-efficiency with other users. I kinda feel like this is a requirement for modern programming games.

7 Billion Humans was mostly fun as far as I played it, and it does boast significantly more puzzles than Human Resource Machine did. But I got bored of its programmer-unfriendly UI and "magic" behavior around the halfway point.

Better than: Opus Magnum, Silicon Zeroes
Not as good as: Exapunks, Human Resource Machine
Nerd alert: if the game supported modular code (functions), I might even have written some type-safe helper modules. Alas.

Progress: 34 "years" (puzzles completed).

Not much to say, here; I tried Metal Gear Rising: Revengeance again and stopped even sooner than the last time I played it.

As I was six years ago, I'm annoyed that the game doesn't do a very good job of explaining what it wants me to do. The difference this time around is that I have less patience to learn how to git gud properly exploit this combat system. How many die-and-retry loops, how many online references, will it take? More than I'm up for.

My more-recent attempt at Transformers: Devastation and replay of Vanquish have also tempered my expectations of Revengeance -- that is, I'm not looking forward to uncovering the rest of Raiden's story. Those other titles reinforced that, past a shiny presentation and bombastic-but-shallow narrative, the "point" of a Platinum game is its combat... and I'm just not that into it.

Progress: Gave up on the Blade Wolf fight in R-01.

Factorio isn't a perfect game -- not even "perfect for me," in the sense that some narrative elements and improvements to tech-tree pacing could make it "more perfect for me." But, well, it's a good thing that I had this past week off of work and could spend multiple full days (and two all-nighters!) building a spaceship factory.

Calling my 60-hour creation a spaceship factory is absurdly reductive, but that reduction is the key to Factorio's charm. To create rocket fuel, you need light oil; for that, you need a chemical plant; for that, you need steel plates... ad infinitum. The layers upon layers of manufacturing objectives, not to mention the required research to unlock them (and the science packs to perform research!), are wonderfully clear in defining "what" you need to do and delightfully vague in "how" to bring it all together.

At their best, discrete components feel like meaningful objectives all on their own, like the firearm magazine which is a necessary staple while also building toward upgraded ammunition. Some components aren't so satisfying, like the aforementioned light oil and its refining counterpart heavy oil. But even then, the fact that every manufacturing problem breaks down into individual parts lends Factorio an addictive "just one more step" quality, which is what kept me playing into the wee hours of the morning.

I called out the game's pacing earlier, and there's one very specifically frustrating part of the tech tree that I'd suggest changing: the full utility of a logistic network (specifically the requester chest) isn't unlocked soon enough. Until you can use drones to ferry material from factory to factory, everything has to be done with conveyor belts, and the complexity of belt layouts required for some multi-stage manufacturing is a damn nightmare.

To highlight this, look at the tree of inputs for the utility science pack, and how many materials are shared by multiple layers. This science pack is required to research the ability to build a logistic network -- so you can't use a logistic network to build it, at least not at first.

Consequently, once I had finally unlocked requester boxes, I spent the better part of a day replacing my byzantine belt networks with flying drones. And then I discovered that I'd exceeded my electricity production, and had to expand my reactor complex.

I'll admit that this manner of "and then..." effect is often torturous; but I also really enjoyed it. Hell, unknown-unknown problem solving is part of what I'm paid to do.

Toward the end, or at least toward where I ended the game (after launching a rocket, you can keep going and build even more insane shit), I would also complain that some building processes just take too long. The raw time involved in making a rocket control unit, of which you need one thousand to launch the rocket, made this part of the game feel like an idler where I spent quite a while sitting and waiting.

I haven't even mentioned Factorio's combat, which ultimately doesn't feel like an important element. At first, defending yourself and your assembly lines from hostile bugs can be downright harrowing, but once you've got the manufacturing capacity to keep up some gun turrets - and once you learn that destroying nearby hives will effectively stop the attacks - combat becomes a very low priority.

Well, and you'll need to keep up an arms race as your increasing pollution levels lead to bigger and badder bugs. But once you can build a tank and load it with explosive shells nothing really stands a chance anymore.

So, yeah, there are parts I was underwhelmed by and parts that I definitely think could have been done better. But none of that changes the fact that Factorio is the first game in years that kept me up literally all night playing.

Better than: Infinifactory
Not as good as: Exapunks
As for what's next: when will Satisfactory leave early access?

Progress: Launched a rocket.

I appreciate that Gotham Knights is being up-front about how it isn't a "Batman" game. Bruce is gone, Bats isn't even part of the title -- and whether or not Knights incorporates any stealth, the footage so far is clearly more focused on combat.

Arkham Asylum was an excellent Batman simulator, while its follow-ups City then Knight gradually added more open-world superheroism. I don't think it's necessarily a bad thing if Knights iterates specifically on the latter, especially since this dev studio has some relevant experience from the Arkham Origins spin-off.

That said, the dialog in the demo trailer... ugh. I would have hoped Batman & Robin was a low narrative bar to clear, but dramatic banter like "We'll see about that" has me nervous.

A lot has changed since I last tried (and subsequently purchased) Factorio: my career, my lifestyle, the state of the world in general, and - critically - my expectations of modern video games. So it's something of a relief that, six years later, Factorio 1.0 still satisfies the same industrial urges as the demo.

I will say that I'm a little disappointed at the lack of narrative complexity; Factorio's setup is more like a Civ "campaign" than a story-driven one. But hoping for anything more was a long shot.

In spite of how popular and bloated the Open World Survival Craft genre has become since 2014, Factorio's focus on logistical automation doesn't feel stale or samey. The game does have a tough balance to strike between big-picture objectives and detail work, due to how many intermediate layers are involved in some recipes -- but at least so far, it feels pretty good. I like standing back and marveling at my well-oiled machinery, and I also like diving into the conveyor belts to fix or optimize specific parts of my supply chain.

What I'm still getting the hang of is the integrate-or-refactor dichotomy. Unlike Zachtronics and other programming-puzzle games, Factorio doesn't just accept your solution and move on; you'll need to continue smelting plates, or assembling widgets, as dependencies of later manufacturing. And while it's tempting to plug output X directly in as an input to Y, this becomes a liability when you discover that you also need X as an input to Z. Sometimes the assembly line for X just needs to be painstakingly re-architected.

I definitely need some more practice to develop good planning instincts, for creating infrastructure that's modular and flexible from the start. ... kind of funny how Factorio's learning curve resembles professional software development in that way.

Progress: Finished the tutorial (after nine hours!).

Architecturally and thematically, Dex is a 2D side-scrolling interpretation of Deus Ex: an open-world cyberpunk dystopia that encourages varied playstyles. It's an ambitious design to ape, and while Dex pulls off a surprisingly-good narrative, the execution of its gameplay leaves much to be desired.

Dex's story starts with a cryptic message from an elite hacker, a "chase" sequence (low-key since it's also a tutorial), and the reveal of the game's central conspiracy: a shadowy techno-oligarchy attempting to control society with weaponized AI. ... that summary may sound a bit hackneyed, but the game's writing and characterization are on-point, full of sordidly detailed technobabble to flesh the story out.

The pacing of story missions is a bit wonky, in part because a lot of exposition is saved for the end, but especially because some earlier missions feel like wild goose chases. To progress the story you need a hacking upgrade; but first you need to get an augmentation from the doctor; but first the doctor needs some equipment from the store; but first the equipment supplier needs you to run an errand; these missions are pretty transparent in their purpose of introducing new characters by stringing you along.

On the flip side, Dex's optional side-missions are more compartmentalized, and really great at making the game world feel alive. These missions are full of well-polished details like emails that reveal characters' secrets, notes that hint at safe combinations... the same kind of satisfying minutiae that makes Deus Ex so engrossing.

Sadly, inbetween talking to characters and hunting for information, the rest of Dex's gameplay is kinda just bad.

Melee combat is a mess, as enemies love to block your blows and knock you down (interrupting the game's flow while you wait for Dex to stand back up). They also take and deal out plenty of damage; I seriously could not get through the game's introduction until I lowered the difficulty. Gunplay is also a mess, as bullets don't do enough damage to stop enemies from getting in your face -- also, the control for aiming is the same as the control for dodge-rolling, so be careful with that!

The hacking interface is basically a twin-stick shooter, and on top of being fairly repetitive, it relies on an energy resource which you have to exit the interface to recharge. Lengthier hacking sequences are an exercise in entering the hack, making incremental progress, going back to the exit, recharging, and repeating several times over.

All of these mechanics can be made less awful by investing skill/upgrade points against them: punch enemies harder, shoot at them more accurately, hack faster, et cetera. But even fully-upgraded, these core components of Dex's gameplay peak at "mildly annoying." At their worst - especially early in the game, when you don't have upgrades yet - combat and hacking are frustrating chores.

And while Dex embraces the spirit of Deus Ex in providing non-combat methods to solve problems, in practice it's got the same flaw as Human Revolution's bosses: there are some fights that you just can't sneak or hack your way around.

Dex is impressively competent at the storytelling aspects of a techno-conspiracy thriller, and its world-building sidequests are genuinely awesome. But the game's non-narrative mechanics are poorly executed and un-fun, ultimately bringing the whole experience down with them.

Better than: Deus Ex: The Fall
Not as good as: Deus Ex: Human Revolution - The Missing Link
This game was "enhanced" at some point in the past few years: which makes me wonder, and fear, how rough its original gameplay might have been.

Progress: Finished on "Casual" combat difficulty.

Ittle Dew 2+ is overall a competent and worthwhile iteration on its predecessor. It's an unabashed Zelda-like with a silly and fun story, and a keen focus on gameplay fundamentals: block puzzles, varied enemies, magic items hidden in dungeons, secret caves with optional upgrades scattered around the world map.

My only real complaint against the first game was that it was short, and this sequel remedies that thoroughly. Not just with a bigger map and more numerous dungeons -- there are also more items and gameplay mechanics to keep that more-expansive content feeling fresh and fun.

There is a dark side to Ittle Dew 2's ambition, however: boss fights. The game's combat leans toward "unforgiving" in general, but bosses are especially heinous in the amount of damage they can both take and dish out.

One dungeon boss had me dying and retrying so many times that I almost broke my chair. I did eventually win, though, motivated by my desire to see my way through the game. And then I got to the eighth dungeon boss, a seriously nine-minute-long encounter where dying at the end makes you restart the whole thing.

Which is exactly what happened to me, causing me to immediately quit the game forever.

Mid-encounter checkpoints, or even just a lower difficulty setting, would have been enough to push me into this fight again -- maybe even to win it, and unlock some post-game content. But the game clearly wants me to suffer here, and to that I say: no, thanks.

Up until that point, Ittle Dew 2 was a true pleasure, and I'm very satisfied with how it improved upon the first game. Exploring the map looking for hidden chambers was a delight, as was solving my way through each dungeon. I've got no regrets about the hours I spent delving into this weird world.

I just won't get to experience the totality of its content, due to punishing boss design. Oh, well.

Better than: Anodyne, Cat Quest, CrossCode, Oceanhorn: Monster of Uncharted Seas
Not as good as: The Legend of Zelda: A Link Between Worlds
On balance, basically just as good as: Ittle Dew; more good stuff, tainted by frustrating bosses.

Progress: Gave up on the Passel fight.

Deadbolt feels like a mildly-Gunpoint twist on Hotline Miami. I love being able to see the level zoomed out, and plan my approach; I love the environmental opportunities to move through air ducts, or flip over a table for cover; I love the flexibility afforded by finding (or purchasing) additional weapons and varying my combat tactics accordingly.

Unfortunately, like Hotline Miami, Deadbolt kills you in one hit and makes you restart the level when you die. I don't love that.

This is a non-issue in the early game, when there are only a couple enemies to get rid of and a mission feels like a short puzzle. But as the levels get progressively larger and more full of instant-kill enemies, the repetition of dying and retrying just gets too annoying.

As much fun as the sneak-and-assassinate action is, and as curious as I am to see new weapons and environments, I don't have enough patience to keep trying a level over and over again.

Better than: Hotline Miami
Not as good as: Mr. Shifty
Maybe the Gunpoint comparison is rough: but the simple side-view art feels similar, and also, I miss Gunpoint.

Progress: Finished a few Zombie Kingz missions.

See No Evil does a respectable job of twisting and augmenting familiar puzzle mechanics. Sure, you push blocks to hold down buttons, and reflect lasers sound waves using angled mirrors bent pipes -- but you also use these tricks to manipulate enemies into hitting unreachable switches, or to distract them away from catching you.

In theory, this is a refreshing combination of sneaking and puzzles (like Boktai, sans combat). In practice, though, See No Evil suffers from some fundamental problems: visual ambiguity and fiddly movement.

The game is shown at an isometric 3/4 overhead perspective, and the art for objects and characters shows depth, but that depth doesn't exactly match up with collision; I frequently activated buttons that I was trying to walk around. Meanwhile, sound wave reflection puzzles require precise positioning while facing in a specific direction, and there's no control to rotate in place -- so getting it exactly right can take several walking-in-circles attempts.

Occasionally these issues combine: trying to push a block around tight obstacles is ... not fun.

As for the story, it's surprising enough that See No Evil has one - this kind of puzzle game tends to skimp - and it's admirably integrated with the game's presentation and mechanics. Between-level exposition scenes shed light on the game's macabre backstory, explaining why your enemies can't see, and why they're resisting your journey; it's kinda neat.

But, not neat enough to goad me into playing more. See No Evil certainly isn't the most frustrating puzzle game I've played, but its precision issues are just frustrating enough to turn me off.

Better than: Great Permutator
Not as good as: Toki Tori (2008)
Definitely not as good as: Mark of the Ninja: Remastered

Progress: Got to Chapter 3.

First thing's first: MO:Astray's Steam tags call it a Metroidvania. It isn't. At least as far as I saw, there's no backtracking, no world map, and not a lot in the way of ability upgrades.

MO:Astray is instead a linear platform-puzzle game with a sci-fi narrative. The story seems fairly predictable: underground science facility, experiments gone wrong, scientists turned into zombies; you know the drill.

The protagonist, Mo, is a blob. It can move left and right on the ground, and can stick to walls or ceilings, but can't climb them. At least, not directly; Mo can jump from one wall to another, which is the foundation of most of Chapter 1's traversal puzzles.

These mechanics are competent enough, but ... kinda boring. The novelty of blob-movement wears off after a few rooms. And while the game does gradually show off additional traversal abilities (like a mid-air jump), these are introduced too slowly to keep the game interesting.

There is one cool twist in how Mo can leap onto a creature's head - be it a zombie scientist or a wall-crawling spike rat - and take control of it. But even this mechanic becomes underwhelming pretty quickly; mind-controlled creatures simply don't do very much.

Meanwhile, the atmospheric ambience can be engrossing in how well it pulls off the "post-incident" vibe, but is just as often spoiled by a general lack of polish. Foregrounds and backgrounds are difficult to discern in a dark room; positioning some movable platforms takes an unwelcome amount of trial and error; a swinging mechanic is inconsistently applied to some objects, but not others.

MO:Astray isn't a bad game, but it's not that compelling, either. There are plenty of other platform-puzzlers, and true Metroidvanias, with more interesting gameplay and storytelling.

Better than: Rise & Shine
Not as good as: The Swapper
Less combat-focused, but arguably on par with: Capsized

Progress: gave up after a few minutes in Chapter 2.