What Games are you playing 2015?

[SnowFire]

SF's mobile gaming:

Fire Emblem Awakening, Equality playthrough: Pretty much got put on hold right after my earlier post due to breaking my 3DS.  Then new 3DSes having a nasty shortage (in the sense of "New 3DS", not a replacement old-3DS model).  I bought a new 3DS last Sunday...  but the memory card is hidden with the battery, and to get to the battery requires taking off the screws, which was freaking impossible for 1 screw.  I'm now sending it back to Nintendo repairs to loosen a screw, this is silly.  So...  no 3DS / DS gaming.  Vita stuff ahoy!

Corpse Party Book of Shadows - Pretty disappointing.  I don't like the general slant the stories seem to be taking and the gameplay doesn't lend itself well to horror to me, at least not without better writing.  On hold.

Castlevania Symphony of the Night, Maria playthrough - So Dracula X Chronicles, despite me owning the old PSP one, was on sale in the PSN store.  Not as much on sale as it should be by Steam standards but whatevs.  Okay I can finally try out Maria...  what the crap is this, it's locked behind finishing an Alucard run?  And SotN itself is locked behind finishing Dracula X?  Sigh.

Dracula X / Rondo of Blood, playthrough #2 - As usual, did pretty good for stages 1-3, got owned on stage 4.  Friggin precise jumps around traps as Richter.  Wasn't really feeling it.

Tactics Ogre: Let Us Neutral Together - Ah strat-RPGs.  Decided it'd be more fun to do a real playthrough than World back to the start.  No Chariot of course (possibly excepting Hanging Gardens knockback BS, we'll see).  Commentary on more interesting levels:

* Prologue: We've taken away Lanselot & his White Knight buddies equipment, let's see if they notice.  Aside from being funny to watch 'em get their butt kicked, it's probably optimal too, more skill points for Denam & Canopus.

* Krysaro: Sorry Leonar, Donnalto died in combat.  Some weird dude with wings shot him with Dark Weight.  It's...  it's a goddamn mystery as to why.  Hand over Felicia for a change of pace in starter clerics pls.

* Arkhaiopolis of Rhime: Damnit Cistina.  Definitely one of the more RNG battles in the game, but thankfully it's also fast.  If the enemy archers & mages decide to focus down Cistina and/or she goes too YOLO and/or the crits & blocks don't go your way, it's possible there's just not a lot you can do here.  Eventually I get rolling and Canopus proves a sufficient distraction.  I tell Cistina some pish-posh about seeking peace, which is why I asked random foreigners to slay my enemies.  We're two of a kind, here.

* Qadriga Fortress: I remember Reiska complaining in chat about the difficulty of this battle, which is fair enough, but I think he was lvl. 5-6.  I was lvl. 4-5 for this, which makes it much more interesting.   It is definitely one of the toughest battles in the game, although going in knowing that and thus being willing to use Blessing Stones helps - I blew 2 of 'em.  (The very expensive revival item.)  Anyway, for those who forget, this is Nybeth & his undead buddies.  Definitely bring 2 Clerics with Exorcism for this one.  Nybeth has great magic defense and Sidestep for arrows, so you really want to dive him with melee...  but I forgot / had Canopus with Bow & shield for this fight, so he can't exactly do it, this requires #YOLO charging in from the frontline.  Which wouldn't be so bad?  Denam is a champion tank at this point, at least with the answers to the Tarot I gave (which generally did focus tankiness), BUT.  It's all narrow staircases up, and corpses block your path.  You cannot walk over the ghost remains...  whatever that means?  Better exorcise 'em first, except by definition this is Denam & Voltare-the-Knight diving deep, so it's not real safe for your Clerics to go that far forward.  I had a positively ridiculous chase around the edge of the fortress as Nybeth retreated.  He had to cast Summon Darkness at least 4-5 times this fight, bringing on new waves of reinforcements nearby, always vexing...   and furthermore these incorporeal ghosts have the ability to, aside from body-blocking your melee units, throw Mend Herbs at Nybeth for a very relevant +25 HP.  Eventually, somehow, after chasing Nybeth from the top-left to the right side of the fortress, I managed to get a Knight to stick to him; in combination with Canopus's Dark Weight earlier (at high TP it'll 100% hit even through Sidestep), I was able to finally take him down, possibly with some chip damage from my mages as well.  Pretty close.

* Balmamusa 2 Chaos: Hmm, C-Denam sucks a lot more at getting out the word than L-Vyce.  I kinda liked how the Duke's plan doesn't really work at dividing the Galgastani in the L-path; the Hireophant's opponents are silenced, not emboldened, kind of like if word of a fishy-sounding operation got out allegedly done by Mossad would probably create more of a "our good name and honor are under attack by bullshit" deal.  Also Denam's dialogue sure makes it sound like he's gonna stop the operation, but I guess he fails despite defeating Leonar's goons.

* Tynemouth Hill: Dang it Arycelle.  Checking a FAQ, I adopt the oracular strategy of handing her a dagger so she'll retreat.  I manage to reduce Vyce to critical just before he could theoretically have chipper-shredded Arycelle.  (He can nearly kill her at full health, and she only had 2/3 health.)

* Krysaro: A Knight boss with Phalanx, thrilling.  Kill everyone else, cue a long 1960s Batman sequence of my characters holding up signs that say "Biff!" and "Whaam!"

* Ndmamsa Fortress: I spare an evil pregnant pirate.  Not something I do every day.  Say what you will about TO, its world is actually pretty reasonable on gender balance, having a pretty wide variety of female characters both in positions of authority and just as random soldiers.

* Golyat: Seriously, a mage boss who charges in?  At least Vyce & the Ninja guy had something approaching durability.  This makes it hard to get more moneyz from murdering everyone and taking their loots.

* Arkhaiopolis of Rhime 2: Xapan cutsceness at like half health, not even critical.  Let me pick up my tarot cards first!

* Almorica: I thought you'd have to kill both Ramidos & Xapan from the description, but nope, just Ramidos.  Good thing too because I was losing pretty horribly and my position wasn't improving thanks to reinforcements.

* Almorica interior: Wait what why does Vyce kill Xapan?  Something about that's not the right way to use a hostage?  Did Xapan try to torture Catiua or something?  Except Vyse is pretty unsympathetic in this scenario, so maybe Xapan was too nice, or making suggestions above his paygrade or something?  Anyway, regardless, Lvl. 11 Vyce slaughtered L10 Ninja Denam.  Dirk +1 did 8 damage, Dirk did 1 damage, Vyce can heal 25 HP in an action and was doing tons of damage to me, even Heart Crusher was only 34 damage.  This is with making a generic character a ninja as soon as possible (but no grinding) to be item-girl in the back, and then only moving Denam over when he was a mere 2 levels behind.  Sheesh, subtractive defense.  Anyway...  switched Denam to an L11 Knight (my only L11 class) and learned Phalanx, which was sufficient (Phalanx whenever Vyce doubleturns the now much slower Denam to avoid getting murdered horribly).  This is much more of a nerfbat war as the two go at it forever and each occasionally tosses a healing item, but for whatever reason Vyce stopped using healing items at some point and let me win.  Got it.

* Decision split: I have to imagine that Duke Ronwey runs his own Pravda or something at this point for his Stalin-esque shifts in favor.  ATTENTION CITIZENS: The Hero of Golyat we were previously hunting as a traitor is actually not a traitor, but the OTHER Hero of Golyat I put in charge of one of my armies was the REAL traitor, so kill him instead.  Also, just how big is Denam's band anyway?  You'd think the loss in prestige of accepting him back in would be greater than any benefit from them.  But then again L-path Denam was able to single-handedly secretly conquer the northern part of the Galgastani Empire so who knows.

Also, this is less related, but apparently N-path locks you out of a pretty large aftergame quest because Gildas is dead (sans World Card abuse)?  Like, shifting fortunes of war, I get it, that's fine, but Gildas doesn't really have much relevance to that aftergame quest, and it's not really a choice on the user's behalf.  Kinda lame, it's fine to lock you out of something if your decision was vaguely related to it (didn't recruit a key person due to screwing up dialogue choices, say), but this is just random.

* Almorica Execution Site: Okay, game.  Dunno what this was supposed to be, but that was a thing.  Although to be too plotty - in the Chaotic & Neutral path, Balbatos's enemies desert him / rise up against him.  Hektor is portrayed as a firm Galgastani loyalist at the execution, though, yet we know from Law-path that he was anything but - he hated Balbatos.  So were there some OTHER bunch of Balbatos-haters who rose up then in this path I guess?!

* Lizardman Execution Site, aka Arycelle Loyalty Grinding: You know, I'm fine with adding an option to keep Arycelle in Neutral path, bizarre and unintuitive as it is, but the loyalty system is kind of shenanigans.  I'd rather there be some kind of scene / dialogue with the bizarre situation now ("Hi, Leonar!  We broke up because you murdered my brother and a bunch of others, but let's let bygones be bygones 'cuz Denam says so") where you have 1 option that keeps her around & 1 option that doesn't, or some bizarre FAQ-baity quest.  "Go have her hack up a bunch of stuff to re-earn her loyalty" doesn't really jibe, nor does the lack of any scene to sell this.  Also I haven't finished Neutral yet, but if any path it would make sense for Leonar to join, it'd seem to be this one...  oh well.  So yeah, I surrounded a lizardman and Arycelle stabbed him a lot while his cleric buddy healed him.

* Cerya Execution Site: what is this oh god oh god did Oz just crit Cerya for like 80% of her HP in one action nooooooooooooo

Time to head off to the Phorampa Wildwood..  I've avoided any grinding in C1 & C2, but that just isn't gonna hold true for Cerya.  The Templars are L12, Oz is L13, my band of L10 & L11ers is not going to threaten them before Cerya loses her head.

Current party: Ninja Denam, Crossbow Canopus, Arycelle, Folcurt, Cistina, Bayin, rando-Enchantress, Terror Knight Voltare (gotta keep TK leveled for Dievold!), Felicia, Mirdyn.  Mirdyn sucks and will probably be canned eventually.  Crossbows on Canopus are interesting - they're less good & fun than Bow & Dagger Canopus IMHO, since you lose the "fly to heights and shoot outside of range" thing bows have, but this is compensated for by some of the best ultimates in the game.  X-Bow Ultimates do huge damage to an *area* in a game where any area damage is rare, let alone good damage.  A change of pace, at least.

[Captain K.]

Codename STEAM: So Henry Flemming, John Henry, the Cowardly Lion and Tiger Lily all took down a giant ice spewing Lovecraftian Horror with a giant cannon on it's head, while Abraham Lincoln insures the safety of Queen Victoria.

...why don't History Books tell us this part of world history?

I Streetpassed Abe Lincoln the other day, that dude really gets around.

[Fenrir]

Verdict: queue up the HLM2 soundtrack and replay HLM1.

This works pretty well. Best tracks for this are Divide, Technoir, Voyager, Roller Mobster, Le Perv, Future Club. Holy shit. Edit: and sexualizer, how could I forget this one?

Binding of Isaac Rebirth: Shit I spent the entire weekend playing this. I beat every boss, and that's about it.
This game's basic gameplay is decent but not that good (Nuclear Throne is certainly ahead in the genre)
On the other hand, this game manages to have endless replay value with:
- Unlocks everywhere by beating specific bosses with specific characters
- A fuckton of items with super cool effects on gameplay that even synergize together to have even cooler, sometimes absurdly overpowered, effects.
Each playthrough you might find another super sweet combo and might unlock another item.

[The Duck]

Verdict: queue up the HLM2 soundtrack and replay HLM1.

This works pretty well. Best tracks for this are Divide, Technoir, Voyager, Roller Mobster, Le Perv, Future Club. Holy shit. Edit: and sexualizer, how could I forget this one?

Binding of Isaac Rebirth: Shit I spent the entire weekend playing this. I beat every boss, and that's about it.
This game's basic gameplay is decent but not that good (Nuclear Throne is certainly ahead in the genre)
On the other hand, this game manages to have endless replay value with:
- Unlocks everywhere by beating specific bosses with specific characters
- A fuckton of items with super cool effects on gameplay that even synergize together to have even cooler, sometimes absurdly overpowered, effects.
Each playthrough you might find another super sweet combo and might unlock another item.

I've been getting more into games like Binding of Isaac and Spelunky (and Nuclear Throne but I only have a few hours in it), which in theory have unlimited replay value.

There was a neat article about people randomizing games like Super Metroid and Zelda where you can customize certain things (having no swords). They also come with some native challenges. It looks really labor intensive but there are a lot of games I'd like this seen done to.

http://www.usgamer.net/articles/route-randomizers

[Luther Lansfeld]

Super Mario Galaxy - I finished this game with about 85 stars. Fun times, the Wiimote controls put it a bit lower than it would be otherwise, but it's a solid entry into the Mario series. Not as good as 3D World though. :p

Paper Mario - Just finished Chapter 3 with the Invincible Tubba Blubba. Why does Tubba Blubba's heart talk? Now I am wandering around a Toy Factory, trying to disrupt the plans of villainous Shy Guys.

[Dark Holy Elf]

Gourry: Neat, thanks for linking that! I do agree with some of the comments on the article that this kind of random design can never replace good human design (Super Metroid is good in no small part because of the way it is laid out cleverly), and I have little interest in games which are always randomised for that reason, but as a spin on a game I've already mastered they do seem potentially interesting.


Castlevania: Order of Ecclesia - Decided to replay it again, on Hard Mode. I'd done Hard Mode once before, but this is still kicking my ass. Up to the Skeleton Cave. Not really doing anything special with this playthrough besides refamiliarising myself with the game after not playing it for 3 years or so, which is too long considering how much I like it.

Fire Emblem: Path of Radiance - Up to Chapter 28! Using mostly a fairly stock team of lots of Paladins and Flyers. Ike, Oscar, Kieran, Astrid, Geoffrey, Marcia, Tanith, Jill, Soren, Calill, Rhys, Mia (who made the team on the back of a good RNG-blessing pre-promotion then promptly didn't gain Strength 12 levels in a row after).

Marathon Infinity - Well I guess I figured I'd finish the Total Carnage series replay I started a while back. This really drove home how this is the weakest game in the series; stage design is extremely scattershot (challenge is incredibly inconsistent; there's a brutally hard finite-oxygen level, Acme Station, which is in the first third of the game and not even the penultimate level which is obviously trying to be more difficult really compares; I actually had to dig up a FAQ once or twice; the final stage is long and strangely dull with pretty much only one enemy type) and the plot just gets too weird and underexplained. I could go into details on that but since I suspect nobody else has played it I don't see the point. <.< Oh well. Currently doing the VidMaster's Challenge levels which are great fun, taking one level from each of the three games (a good one, in all cases) and remixing it into a nasty challenge level. That will probably scrub some of the bad taste from my mouth, at least!

[The Duck]

Right, one of the comments mentioned a randomized Dark Souls, but the big draw of that series is the obvious care that goes into each encounter and location. In the smaller indie games that use procedural generation, the common thread is that they have relatively simple gameplay conceits that have enough items/powers to provide a varied experience when things get randomized. Perhaps there could be a deliberatively designed Binding of Isaac for example that would make for an awesome, balanced experience, but whatever algorithms are generally good enough to create interesting experiences each playthrough, although you can still sort of get screwed on a crappy world. I hadn't come across the idea of doing that to an established game until reading that article but that opens up a ton of interesting possibilities (again, so long as someone is willing to do the considerable legwork to get there).

[Jo'ou Ranbu]

Donkey Kong Country Returns 3D - World 6, only Perilous Passage and the boss left to go. Every time I start thinking "hey, the game is going easy on me, I'm beating stages losing only one or two lives", they throw me a curveball. The temple stages in particular are simply -insane-. I kinda dislike how geared towards dismantling momentum the physics can be, but I -do- like the insanity scale they apply to the level designs (randomly crumbling canyons left and right? City-sized locust swarm chasing your ass? Riding a minecart inside a giant rolling dinosaur egg? Sure, why not).

But I honestly feel the stage design's also at its absolute -best- when on rails, I feel, mainly because you don't have to fight against Donkey's oddly jerky maneuverability. It's quite a departure from DKC2's fluid momentum pickup and makes me wonder if Time Attack wouldn't be more painful than fun. The stage design itself often indulges in momentum breaking, which also strikes me as strange. It's one of the things that made DKC3 stage design such a chore, though DKCR doesn't pull that anywhere near as often as DKC3 did. Still, it's an interesting and amusingly brutal platforming romp so far.

[Meeplelard]

Donkey Kong Country Returns 3D - World 6, only Perilous Passage and the boss left to go. Every time I start thinking "hey, the game is going easy on me, I'm beating stages losing only one or two lives", they throw me a curveball. The temple stages in particular are simply -insane-. I kinda dislike how geared towards dismantling momentum the physics can be, but I -do- like the insanity scale they apply to the level designs (randomly crumbling canyons left and right? City-sized locust swarm chasing your ass? Riding a minecart inside a giant rolling dinosaur egg? Sure, why not).

But I honestly feel the stage design's also at its absolute -best- when on rails, I feel, mainly because you don't have to fight against Donkey's oddly jerky maneuverability. It's quite a departure from DKC2's fluid momentum pickup and makes me wonder if Time Attack wouldn't be more painful than fun. The stage design itself often indulges in momentum breaking, which also strikes me as strange. It's one of the things that made DKC3 stage design such a chore, though DKCR doesn't pull that anywhere near as often as DKC3 did. Still, it's an interesting and amusingly brutal platforming romp so far.

And to think, the sequel, Tropical Freeze, is basically more of that +5!

[Sierra]

"Procedurally generated" is a phrase that usually kills my interest in a game. I try to remember, "Hey, it worked for Toejam & Earl," but no, that's just not enough.

[Fenrir]

Procedurally generated has come a long way recently. People are more interested in making the perfect algorithms for the best experience vs just having stuff be random for variation. I don't think Toejam & Earl is a particularly good example of this, as the basic gameplay isn't good enough on its own, and luck is a huge factor.
I'm really curious to see how the randomized levels in Bloodborne will end up, but I'm not holding my breath. I think level design is too important for the Souls formula. Flat rooms with 3 Taurus Demons don't sound too good. Whether randomization can be better than human crafted levels depends entirely on the genre and importance of atmosphere. As a pure gameplay experience Super Metroid is a hell of a lot worse than the best new action roguelikes (even on the first playthrough) but it's trying to do something different.

I find that Binding of Isaac is an entirely different beast than Spelunky / Nuclear Throne / Necrodancer / Teleglitch type. Those are extreme skill based challenges that stay relatively similar every time with luck being a medium factor. Randomization makes every game feel different but doesn't radically change the game.

BoI is more about watching really cool fireworks and see how they affect gameplay, with some absurdly broken combos you'll never see again. The challenge is less skill based, the game's overall easier (I win about half the time if I'm using a good character like Eve, about all the time if I'm using Azazel), there are more cheap hits but health is higher. If you get lucky you win, if you don't, eh you still have a pretty good chance.

The Lost is absolutely terrible design for BoI (one hit and you're dead) even though it could work in NT / Spelunky as an extreme challenge. It's already in Necrodancer and it's fine there. (I'm absolutely not touching it though, I'm not good enough)

[NotMiki]

Gourry: Neat, thanks for linking that! I do agree with some of the comments on the article that this kind of random design can never replace good human design (Super Metroid is good in no small part because of the way it is laid out cleverly)

Games where lack of upgrades lock you out of places you need to go can produce a really interesting experience, but in general they're exactly the wrong type of game to try to randomize.  Need that one thing to go that one place?  GFL.  The right type of game is one where there's a variety of powerups that significantly improve your character's performance, but you're never locked out of anything.  A game that pushes you into dealing with various challenges in different ways depending on what you've got on hand.  That type of game rewards ingenuity and discretion.

I digress.  Play Risk of Rain.  If you like randomized elements and platforming, I guarantee you will like it and if you don't that means you're wrong.  Risk of Rain's levels are lightly randomized, though many elements are always present.  Most of what is random is the powerups you get, and those can change the entire flavor of your experience.  To elaborate, you pick a character class, each class has four active abilities, you can carry one active item in addition to that.  Throughout the game there a ton of items will drop, the vast majority providing passive benefits: 10% chance to fire a homing missile on hit, heal while stationary, enemies within a certain distance of you take damage over time, chance enemies explode when killed, etc.  Items work together to produce a beautiful chaos.  A successful playthrough takes 60-90 minutes.

[Fenrir]

I digress.  Play Risk of Rain.  If you like randomized elements and platforming, I guarantee you will like it and if you don't that means you're wrong. 

Yeah I didn't like it. This felt like a game that only becomes fun after you know the items well, 20 hours in, as combat felt a bit weak to me and platforming didn't really mesh very well with it. I'd rank it the worst out of all these quasi roguelikes I played (Nuclear Throne > Spelunky > Crypt of the Necrodancer > Binding of Isaac >> Teleglitch > FTL > Risk of Rain. I should try out Sunless Sea later)

[Grefter]

Gourry: Neat, thanks for linking that! I do agree with some of the comments on the article that this kind of random design can never replace good human design (Super Metroid is good in no small part because of the way it is laid out cleverly)

Games where lack of upgrades lock you out of places you need to go can produce a really interesting experience, but in general they're exactly the wrong type of game to try to randomize.  Need that one thing to go that one place?  GFL.  The right type of game is one where there's a variety of powerups that significantly improve your character's performance, but you're never locked out of anything.  A game that pushes you into dealing with various challenges in different ways depending on what you've got on hand.  That type of game rewards ingenuity and discretion.

That is really the key to improving your procedural generation, ensuring that the generated output is winnable.  It is a case of ensuring that certain things are acheivable (It can just be ensuring that a sequence of things are available in a series of gatekeeper flags? Then in case of failure regenerate world?  That is not the most efficient way to handle it during generation, better to build in to your generator, but harder and "less random" I guess).




Buuuut you hit on the reason these games are so appealing there Elf.  Running them through games you have already "mastered" is interesting.  Generally the appeal is that part but you need to "Master" them while handling the RNG elements. 

BoI actually introduced the ability to pass around the I have hopes that there are good seeds that can be introduced as sort of a "canon" beginners guide run so you can retroactively have a "normal" experience that people can progressively get better at.

These kinds of games really get a weird meta level to them where high level play becomes about managing risk inherent to the procedural generation system based on systems that are actually entirely outside the actual play loop of the game itself.  Fascinating stuff and I hope Alex does a drive by post on this sort of thing because I personally think he is probably the person in the DL that has a ton of experience with it and ability to verbalise it well.



Necrodancer IS the best, but only because it is the only roguelike with a dubstep soundtrack to it (because that is the best thing to play Necrodancer to).

[Twilkitri]

Jurassic Park: The Game - beaten

Entertaining enough. Nothing really new to say about the game that wasn't said in my original report.

Final totals: 25 gold medals, 11 silver medals, 8 bronze medals, 1 lost medal, and 21 deaths.



Citizens Of Earth 3DS - beaten

Regarding the zoom thing - apparently its main use was as a placebo, as I had begun having less trouble evading enemies after zooming out to the limited extent available from the untrained citizen... but after finally recruiting them (and one of the subrequirements for that is in a very late-game area, depressingly) and training them up, I found that the game actually resets the zoom level whenever you move to a different map. Presumably the difference was little enough originally the first time that I just didn't notice it happen. Of course, given that it works this way it means that the zoom option is next-to-useless to use when you haven't recruited the citizen yet because in addition to the restrictions on how far you can zoom, you can only actually have it in effect in the handful of areas that the citizen appears in naturally.

I did end up zooming completely out in some late-game areas which are frustratingly large, but the majority of the time I didn't bother setting it back up.

Another late-game citizen lets you change the game speed, and that setting persists across maps (thankfully). Possibly the zoom level is supposed to persist across maps but there's a bug somewhere, since there's at least one map which has a different default zoom level.

With regards to changing the game speed, it's kind of annoying to have to deal with the reality that the game has been generally slower than preferable but now you can change it to a more reasonable speed and it runs fine without any issues. You can also change it to an unreasonable speed but then you can expect possible issues (mainly walking through things you shouldn't be able to walk through). If you're insane, you can also change it to a slower speed.

More complaints:

• Got stuck in geometry several more times. Most of the time I was not able to escape and had to reload (reminder that this requires closing and restarting the game).
• The game crashed at least ten more times. Three of these were harder than usual in the sense that the 3DS didn't wake up and kill the application, and the home button was unresponsive - holding the power button down still worked, at least. For reference, two of those happened while ambushing a 'Mr. Maze' enemy. Also, the game softly crashed twice while entering fights with the final enemy. Fun fun fun.
• Minor point, but enemy resistances and weaknesses aren't recorded in the Almanac.
• When you have bridges built, the overview maps aren't updated to include them, which is unhelpful at best.
• The Lifeguard broke in some way and started covering for near-death enemies, even from attacks she was making herself.

I ended up beating bosses from some later chapters before one from an earlier chapter, which broke something somewhere that ultimately led to it being impossible for me to trigger the fight with the earlier boss. Didn't prevent me from finishing the game, pretty much just led to uncompleteable tasks on the checklist, so how large of a problem this is probably depends on the observer.

Despite everything, I did enjoy completing sidequests and seeing the different citizens and what sort of attacks they had and so on. Depressing that it has to be so broken - not to say that it would be good if all the bugs were fixed, there are issues with the storyline being almost non-existent for a large proportion of the game, and I don't care much for the battle system, and the fast travel options could do with being considerably improved, and late-game maps tend to be far too large with far too many enemies, and...

I feel like the cast would make for an interesting stat topic (setting aside issues regarding how bonus stat points should be distributed) but I'm certainly not going to do one. Rank VP's Car.

[Sierra]

Procedurally generated has come a long way recently. People are more interested in making the perfect algorithms for the best experience vs just having stuff be random for variation. I don't think Toejam & Earl is a particularly good example of this, as the basic gameplay isn't good enough on its own, and luck is a huge factor.
I'm really curious to see how the randomized levels in Bloodborne will end up, but I'm not holding my breath. I think level design is too important for the Souls formula. Flat rooms with 3 Taurus Demons don't sound too good. Whether randomization can be better than human crafted levels depends entirely on the genre and importance of atmosphere. As a pure gameplay experience Super Metroid is a hell of a lot worse than the best new action roguelikes (even on the first playthrough) but it's trying to do something different.

I find that Binding of Isaac is an entirely different beast than Spelunky / Nuclear Throne / Necrodancer / Teleglitch type. Those are extreme skill based challenges that stay relatively similar every time with luck being a medium factor. Randomization makes every game feel different but doesn't radically change the game.

May be related that a lot of what you're citing sound to be roguelike games, which are something I already dislike on a conceptual basis. I don't think a high degree of randomization really hits on what appeals to me about games.

Touhou Labyrinths: chipping away at postgame still, no major progress to report but am just now noticing that Wriggle's class is officially listed as "A Bug of Light Wriggling in the Dark." And that's just outstanding.

[NotMiki]

Gourry: Neat, thanks for linking that! I do agree with some of the comments on the article that this kind of random design can never replace good human design (Super Metroid is good in no small part because of the way it is laid out cleverly)

Games where lack of upgrades lock you out of places you need to go can produce a really interesting experience, but in general they're exactly the wrong type of game to try to randomize.  Need that one thing to go that one place?  GFL.  The right type of game is one where there's a variety of powerups that significantly improve your character's performance, but you're never locked out of anything.  A game that pushes you into dealing with various challenges in different ways depending on what you've got on hand.  That type of game rewards ingenuity and discretion.

That is really the key to improving your procedural generation, ensuring that the generated output is winnable.  It is a case of ensuring that certain things are acheivable (It can just be ensuring that a sequence of things are available in a series of gatekeeper flags? Then in case of failure regenerate world?  That is not the most efficient way to handle it during generation, better to build in to your generator, but harder and "less random" I guess).

Even assuming you create a generation algorithm that never locks you out of places, I'm skeptical about the Metroid sorts of games working out in practice.  A lot of the challenge in Metroid is in figuring out where to use your new powerups.  Even if you don't know the layout, if you understand the solution to each potential problem, the game is still going to be reduced to a rather pedestrian "find A to open A₁" experience.  Sorry, guys, you already beat Super Metroid and there's nothing you can do to un-beat it.  I think combat-heavy games are a better fit for randomization because unlike games where the main challenge is exploration, combat almost never straight-up locks you out of a path.  It may make a path extremely difficult, but grappling with that fight-or-flight moment is what makes randomized games fun.

[Grefter]

Yeah I was going to comment that there is a reason it is more invested in Roguelikes because the play in Roguelikes tends to be simple enough that it plays well into those mechanics.

Like the first thing I noticed about BoI is that it feels like a 2d Zelda dungeon (as a "Roguelike" with the procedural generation).  How do you port Zelda dungeons to procedurally generated stuff?  Strip out tons of the gatekeeper elements to traversal power ups.  Make the literal key power ups only ever lock out side areas.  Which functionally is paring back the Zelda gameplay so that it works just on the level of a Rogue.

So yeah you could do it with "Metroid" but in that process without even more refinement I do think it would lose a bunch of what makes Metroid games part of that "Metroidvania" genre.

[Meeplelard]

Shovel Knight: Replayed this game on a whim.  Took me 15 seconds more than my previous file, so I can safely say the game should take over 4 hours to complete.  On the upside, I got more music Sheets this time around, found all but one relic in the proper stage (meaning buy at reduced price on top of, you know, being able to use it half a stage earlier), and even managed to buy the Gold Armor, which is really the MOST IMPORTANT ARMOR IN THE GAME!

It makes your armor sparkle and lets you do backflips, having absolutely no gameplay benefit whatsoever and costs 10k to make.

[SnowFire]

Tactics Ogre: Cerya Must Die: This mission is sadistic.  7 Cerya deaths so far.  I can entirely see why it's considered the hardest battle in Tactics Ogre.  Even powerleveling doesn't do very much for you (although some certainly recommended it).  So, some details.  Oz & the Templars start at the top of Boed Fortress, you start at the bottom.  So your archers are way less OP than usual, they certainly won't be sniping the squishy backline mages & clerics who will stay high & safe.

* The AI is screwy.  To give Cerya a fighting chance, Oz always passes his first turn and the regular fighters seem strangely disinterested in going for Cerya.  Fine.  Except... the mages will certainly go for her, and Oz will start cracking skulls with a 2HKO to Cerya quite quickly, so you need to hustle it on over.  Healing magic is pretty weak, though; it takes ~2 heals to cancel a single Oz attack.  This means that once the mages start throwing spells at her, it's really hard to keep her afloat; and even if she has full health, Oz attack + 2 mage turns in a row can just plain KO her from full, or nearly so, and you can't guarantee that won't happen even with Slowing Oz.  I might know this from bitter experience when I thought Cerya was safe, but the turns lined up wrong.  Items would help, but...

* ...Cerya usually gets surrounded by 2 Templars to a corner such that there's no room to have Canopus stand next to her and heal.  And the vertical tolerance is way too high to heal from below, or Lobber from below.  So items are mostly out.

* I have seen the Templars blocking Cerya move out before, though.  She used the opportunity to go punch Oz in the face and eat a crit damage counter for 60 or so.  but hey she got 1 damage in so worth right??

* And as a reminder, defensive buffs in TO really suck.  Fortify is like -3 damage a hit or something puny.  There's apparently some aftergame Draconic Magic that will null the next attack but hahahano for having that now.  So don't think about buffing up Cerya.

* Congratulations!  You've somehow, through a combination of AI luck and using Boon of Swiftness, manage to hustle it up to heal Cerya to not die instantly.  Maybe you even killed the most forward Terror Knight, and perhaps the archers raining shots on you didn't get a crit-knockback on your cleric to knock her off the path upward.  However.  Oz has been gaining mad TP from hitting Cerya in the face.  He's got good defense, but if you go in on him, he's just going to hit 100 TP.  And if you don't attack him, he'll, uh, still get 100 TP pretty quick, albeit maybe a turn slower.  And yeah, even when I gave him other juicy targets to hit, he likes Axe ultimating Cerya with Dark Prison.  This isn't quite a OHKO but if Cerya is at anything other than full health she will die, and keeping her at even half health is quite the struggle.  (This is also why being overlevel won't necessarily help - you'd need massive damage to kill Oz so quickly he doesn't get a turn at 100 TP.  Your own ultimates would help but you won't have built up to them before he gets his.)  And yes, Dark Prison has range, so if Oz walked away from Cerya, he can likely still do this to her anyway.

So...  what's the plan?  My own personal thoughts - confirmed that somebody else solved it this way- is something like this: replay the battle until somehow there gets to be some distance for someone to sneak in next to Cerya or Oz.  Don't ask how.  Then get a Knockback-powered blast to knock either Oz or Cerya off the ledge - ideally Cerya, but this only works if she's in good health.  If Oz is knocked down, then just murder him; if Cerya is knocked down, she should be at an angle that makes the mages unlikely to threaten her at least.  I have no idea how I'm going to induce this, but we'll see.

Siiigh.

[Jo'ou Ranbu]

If buffing defenses would help, would petrifying Cerya and then having people heal her work? I don't remember how potent Petrify's damage reduction is, though, and that also presents the problem of having her end up as even more of a sitting duck.

[Luther Lansfeld]

I have never done Neutral in PSP remake, but in the original I pretty much just determined that it was impossible to save her and moved on. >_>

[jsh357]

I want to say I used the Teleport spell to save her in the SNES version, but it's been so long that I couldn't say.  Not like it's worth doing, at least.  To clarify, I mean teleporting one of my units up to go clobber the enemies.

[SnowFire]

Jo'ou: Stone probably would help!  I'm too low level to learn it, though.  Dark magic Petriburst is L13, Earth Magic Petrifog is...  more, I forget.  I was previously at L10-L11; I've leveled my team up to L12 for the fight to be more sane.  It wouldn't be unreasonable to hit L13 though to give an extra push to my offense, and if I do I'll give petrifying Cerya a shot.

Ciato: Heh.  Unfortunately I happen to know that there's an event / battle / more recruits or something that happen in C4 if she's saved in C3, so I do want to give a good shot to saving her.  If it was purely a matter of pride a la winning S3 optional fights and she was guaranteed to live/die regardless of the battle, with just different dialogue, I'd definitely have given up by now.

jsh: Teleport is still in PSP, but it's aftergame Draconic magic, so not working here alas.   One idea from the Internet I didn't have before is to use Wheel of Fortune tarots to give my ground units Blinkwalk (ignore height when moving), which gets them up the fortress a good deal faster - but I only have 2 of them.  Maybe tarot drops are farmable via Chariot?!

[Grefter]

I am kind of tempted to replay just to see that scenario.  If Ignore Height is the answer then wassname crossbow Hawkman would be my knee jerk fix for the problem, but clearly that isn't the case here.  That means that is a crazy scenario.

But fuuuuuck replaying that mess of a crafting system that hits me weak points for massive damage.