Back in the StarCraft and Brood War days, the backstory barely intruded into the game at all, and it's hard not to miss that here. The narrative between missions in [Heart of the Swarm], more often than not, feels like wasteful filler.

Has it really been... huh, yeah, almost three years. And as for Legacy of the Void, or at least its "Whispers of Oblivion" mini-campaign preview, it looks like more of the same.

Wings of Liberty and Heart of the Swarm turned Zeratul into a complete MacGuffin, appearing out of thin air to tease Raynor about the xel'naga or the hybrid threat, and - contrary to the other characters' flat personalities - never presenting any personality at all. If Legacy of the Void is to be his campaign, I dread a story that isn't just embarrassing, but outright dull.

But that probably won't stop me from playing it. Even though the mini-campaign's three missions didn't diverge very much from established StarCraft campaign mechanics (going so far as to combine a small base-building mission with an infiltrate-the-laboratory mission). It's still good fun. And if I can take out narrative enemies by massing Void Rays like an asshole, then, hell yeah.

Progress: Finished the Whispers of Oblivion prologue.

This has been "in the works" (barely) for a while, but I've finally transitioned the Glog to a fully static site. I'm now using the Hugo site generator, with a personal modification of the Greyshade theme, to render and serve Glog pages completely statically.

Why the move away from WordPress? Well, aside from the ever-present risk of shitty PHP exploits, I just got sick of running software upgrades on the site, and its plugins, that required me to re-apply Glog customizations. Now I can write a post in my text editor, hit a button to publish it online, and never have to worry about the site again until I decide that I want to.

Editing my Hugo theme is a lot easier than screwing around with a PHP theme in WordPress, too, although it still has the same basic problem of not making it very clear what templates/functions live in the core framework vs. what lives in the theme customizations.

(I had begun a plan to write my own static site generator, instead of using something pre-baked like Hugo, but when I realized that I'd forgotten to implement list/archive pages ... I gave up. Maybe I'll try again in another two years.)

All the old content should be here, and all old posts should even be accessible at their original URLs. Paths for some game-specific post lists have changed, though, due to differences between how Hugo and WordPress make URL-safe strings.

Of course, as a fully static site, there is no longer any such thing as a comment or user account here. Even better!

There are some great things that Grey Goo does really well, like well-differentiated units, interesting mission objectives, an intense soundtrack, and an aesthetically-strong campaign, with gorgeous cinematics and believable storytelling (at least as far as I saw). Unfortunately it stumbles on some really fundamental stuff that ruins the game, even after I feel like I should "get" it.

It's not just that the basic hub-and-spoke building mechanic is unintuitive; it is, but that's not the whole story. Even after I learned that I had to keep factory F connected to hub H and attachments A1 and A2 also connected to hub H in order to build units X and Y with factory F -- maintaining that mess was a chore. Having to remember which factories can build which units is a silly problem to have, when other games will simply have different buildings for different unit types.

But even that wasn't bad enough to make me quit the game in frustration. The problem I couldn't take any more of was hub placement.

Different sizes of hub buildings can have different numbers/arrangements of other buildings connected to them, and the way in which the map allows or disallows particular connection positions is absolute fucking voodoo. There appears to be an effect from terrain height, as implied by grass density on the map?, but not all the time. And although you can rotate a hub while placing it, you can't rotate it until after a location has been picked -- so, if none of the possible rotations work out, you have to start at a new pixel-perfect location with the default rotation, and try everything all over again. I wasted minutes at a time trying to place a medium hub such that it could actually use all four of its possible connections, and still failed most of the time.

When the hardest thing about a strategy game is figuring out where and how to place the most basic structure, then the game is wrong.

Progress: Gave up in mission 3.

A well-traveled gamer might observe CastleStorm's hodge-podge of numerous mechanics and inspirations, and suspect that much of the game is undercooked. This suspicion would be pretty accurate. But to end the story there would be a disservice to what the game does right.

CastleStorm is primarily a 2D, side-scrolling real-time strategy game with three interwoven threads of gameplay. One is a Swords & Soldiers-style marching war, where units spawned by the player march up against units spawned by the enemy -- with different unit types having different strengths and weaknesses. Archers can deal considerable damage but are easy to dispatch, paladins are strong but slow, et cetera. The goal of these troops is to march into the enemy's castle, steal their flag, and take it back to the player's castle.

These units can't be controlled directly once they're spawned; but another thread to the game is temporarily spawning and controlling the "hero unit," a super-strong leader who can move and sword-swing and arrow-shoot on-demand. While the hero unit can't pick up the enemy flag, he can handily clear the way for ground troops.

The third gameplay thread - frankly, the interesting one - is siege weaponry. From the invincible turret at the front of the player's castle, various projectiles can be aimed either at enemy troops or at the enemy castle. Castle damage can cripple the enemy by destroying unit-producing facilities, or the game can be won by obliterating the enemy castle altogether.

So, castle destruction is clearly an awesome thing. And when it works, it works well. But CastleStorm trips over itself a few times on the way to delivering this fine, castle-exploding experience.

Inbetween levels of the campaign, units and weapons can be upgraded (using money-points earned in the levels themselves). There are too many things to upgrade, and the game occasionally - especially in earlier levels - restricts the use of troops, or the hero, or the turret; resulting in many hard-earned upgrades going to waste. As the campaign proceeds, more troop and weapon types are unlocked, making some upgraded units obsolete. But there's no way to recover the money that's already been spent on old units.

Another between-levels activity is building the castle, which could have been a quaint nicety but instead feels like a boondoggle of a feature. Castle customization is unavoidable -- unit spawners, and gameplay bonus rooms, must be placed in the castle in order to have any effect. But the UI for selecting and placing rooms is an irritating mess. Moving rooms is the worst, since it is likely to require deleting other rooms in order to make space, only to re-create them in some new arrangement.

It's disappointing that the full-blown castle designer is required at all, since there are really only two requirements for any given castle:

• To include the rooms needed for selected units and bonuses;
• And to shield them from siege attacks by placing other rooms in front.

The option for cosmetic castle arrangement is ... neat, I guess, but using it to add a new room to a standard castle design feels like loading up Microsoft Office just to type a sentence.

And there are enough other miscellaneous aggravations about the gameplay - friendly fire damage from the turret, wonky camera zoom behavior, being unable to launch new attacks while waiting for a bomb to explode - that it's hard to be satisfied overall with CastleStorm. From concept to execution, much of the game is simply lacking in release-quality polish.

And yet! CastleStorm ultimately succeeds in being more than just the sum of its parts, thanks somewhat to the smart design of its campaign levels. Level objectives are mixed up with a healthy amount of variety, and most levels are short enough to keep the sequence of unlocking and upgrading moving along at a brisk pace. There is even a bit of humor in the campaign's storytelling, although the narrative is just barely deep enough to explain why two castles are fighting each other.

That's why, despite my misgivings toward CastleStorm, my "hey let's try this out" first sitting with the game turned into a 3+ hour journey through the entire main campaign. CastleStorm nails the addictive qualities it needs to keep going, regardless of its other annoyances.

That having been said, I don't think I will be coming back to it for the additional campaigns. At this point I feel like I have wrung all I am likely to enjoy (and then some) out of CastleStorm.

Better than: Swords & Soldiers
Not as good as: Kingdom Rush
Would seriously have enjoyed it much more: If castle room placement was replaced with a list of checkboxes.

Progress: Completed the first campaign on Normal.

Oceanhorn is certainly a Zelda clone -- there's no doubt about that. But it's not quite the clone you might assume at first glance. While Oceanhorn immediately evokes the design and aesthetics of Wind Waker, it actually plays out much more like Link's Awakening with an island-hopping twist.

The story is eerie and curious (up until the explanation-heavy ending). The mechanics are two-dimensional, plus some terrain elevation. You can equip one item at a time -- one of which allows you to jump, basically. Not to mention heart containers, a magic meter, small keys and boss keys, sword and shield, yadda yadda yadda. Oceanhorn does a really admirable job of emulating some of the best parts of a Zelda game, even if it's fairly short by comparison.

Which makes it both baffling and disappointing when Oceanhorn does something really, really wrong. Like extremely heavy auto-aim on ranged items, making it practically impossible to target the right thing in tight quarters. Or bombable walls and obstacles that have no visual indication they're bombable, other than being in a suspicious dead end. Or the great sea, on which you merely choose your island destination and play a brief on-rails shooting game to sail there.

But Oceanhorn's highest crimes are, unfortunately, technical. I suffered very frequent graphical glitches, with terrain and enemies and the protagonist disappearing altogether, requiring me to flip switches in the graphics options just to reset rendering state. The game is capable of switching so rapidly between beautiful and unplayable, it's just ... frustrating. And while there were only a couple of boss fights where the over-the-shoulder camera (as opposed to its normal isometric 3/4 view) got stuck in the map, it got stuck a lot.

Oceanhorn is a great idea, with some cool execution, tragically marred by some glaring faults. What it does well is worth playing, but not really worth suffering through the entire game for.

Better than: Evoland (particularly the Zelda-like parts of Evoland)
Not as good as: Ittle Dew
Hard to say between this and: Anodyne

Progress: Completed the story, 79% items

With how hard and fast the hype train for Phantom Pain has been running, I'm glad I finally tried out the Ground Zeroes "demo," first; saves me some consternation about whether an open-world Metal Gear Solid would finally do it for me. (It doesn't.)

Contiguous open-endedness is a big plus for me, and unlike - say - Hitman, Ground Zeroes does a lot to avoid strict and obtuse punishment for being detected; if I fuck up and get spotted, running to cover is almost always an option. But the gunplay is unsatisfying enough to make it clear that this is not the way Kojima wants you to play. And unfortunately, when you're focused on sneaking around, large open maps become easy to get completely lost in; I spent most of my playing time slowly covering the same ground, over and over, unable to find the next mission objective.

And of course, then there are the overwrought cutscenes. Ground Zeroes is not a long "game" (having only one primary mission), and the longest part of it was when I was wandering aimlessly looking for the final objective -- but the second longest part was the interminable stream of ending cinematics. Many modern action games are rightly criticized for reducing high-octane story sequences to barely-interactive quick time events, but at least that's better than just sitting and watching your playable character do a bunch of cool shit. Every time the camera panned behind Snake during the ending battle, I hoped I would be able to control it, and my hopes were dashed each time.

I can see the appeal, I guess. Metal Gear Solid feels like it's become the video game version of Tom Clancy-style military nerddom. If you're into the realistic hardware and the fantastical-but-politically-grounded plot, then, great. I'm not, and the gameplay isn't enough to carry itself for me.

Valiant Hearts really feels like it does "good enough" on all fronts. Despite occasional frustrations with pixel-hunting or unclear objectives, the adventure-style gameplay is solid and engaging. Despite the relatively generic characters, they manage to eke out some emotional moments toward the end. And despite both of the aforementioned angles, the game also manages to present powerful and educational insights into what it was like to live during - or through - World War I.

It isn't a revolutionary triumph, but Valiant Hearts accomplishes an admirable amount for what it is.

Better than: Monkey Island 2 Special Edition: LeChuck's Revenge
Not as good as: Broken Age (Chapter 1, anyway)
If it helps, might be worth mentioning: That there are a handful of pseudo-runner, obstacle-dodging stages set to era-appropriate war and marching music.

Progress: Completed the story, missing some collectibles.

Arkham Knight looks gorgeous running at 60 frames per second. It's a shame that the game doesn't work at that speed.

But aside from the easy target of Arkham Knight's technical infidelity - which, by the way: I had to restart the Riddler boss fight because he got stuck in an AI loop - it's just as disappointing that the game's most lackluster sidequest is also its largest and most time-consuming. The Riddler's collectible bullshit accounts for almost half of my total playing time.

While I can admire the cleverness of many of the Riddler's puzzles, the sheer amount of them makes dealing with him more tedious and annoying than anything. Which, I guess, actually fits The Riddler's character rather well.

Progress: 100%

I'll always remember Iwata-san as an engineer-at-heart who helped bring joy to the world.

As much as I enjoy getting yelled at by Dr. Venture - and, honestly, it's way more than I expected to - The Magic Circle really failed to bring me around. It feels more than a little too "meta" when a game lampooning incomplete game design has, itself, highly linearized and stilted storytelling content. The fiction behind TMC's fictional production process has some cool mystery to it at first, but once some of its beans start to get spilled, the mystique really wears off.

The Magic Circle's gameplay mechanics are not nuanced enough to make a game out of; really, this feels like a humorous story told semi-interactively. And within the bounds of the burgeoning "walking simulator" genre, what makes or breaks this kind of storytelling is how interesting the world is. Gone Home, from its opening moment, used ambient drama and detailed world design to make you want to know what it was about. What I played of The Magic Circle reminded me more of Dear Esther, where the environment you walk around in exists purely for its own sake; to fill time inbetween scripted story beats.

That this story is a well-written satire of an absurd creative process is ... ironic? I think? But as in Matt Hazard, there's an important lesson here about how easy it is to make fun of poor development practices, versus how hard it is to execute good development practices.

Maybe I'm missing some critical information; but Steam reviews suggest that the full game is about four hours long, so if the first 10-15% of that isn't demonstrative then I would be somewhat less than impressed anyway.

Progress: About 30 minutes into the demo.