Roll Eyes
Lips Sealed
123 ->
List results:
Search options:
Use \ before commas in usernames
So, I'm working on a single-segment run of Dragon Warrior VII, that crazy-long PS1 game.  Specifically, the US version.  (The Japanese version is probably an hour or more shorter just from less text, but I don't own and have never played it.)

I couldn't find any speedruns of this game, not even a TAS.  (Actually, I've heard that the game's RNG is rigged, so it wouldn't surprise me if there was no TAS.)  Instead, I've developed the whole strategy on my own.  This is basically my first speedrun, but I've been working hard on this.

There's some grinding involved in my current strategy.  Some of it's just to make things safer, but I often just need particular abilities to speed up exploration, or to survive battles.  I have fairly consistent strategies for every major battle, as well.

The battle mechanics and story triggers are insanely convoluted, but I think I have it all worked out.  I've also written out a huge list of items that I pick up, often just to sell.  Equipment is incredibly helpful in this game.

I believe I can consistently complete this game in about 21 hours.  (And yes, I think that's actually pretty quick for this monster of a game.)

Under 20 hours should absolutely be possible if additional risks are taken.  But given the sheer length of the game, I just want to get through it without any deaths, first.  Dying just once can cost up to 30 minutes.

I haven't actually attempted a single-segment yet.  But this is a 20+ hour game, and just playing for that long would take some setup.  I'd need a second person to help me, for one.  I also don't have any means to record it yet, either, and I'm not about to try a single-segment without recording it.  (I'm working on that, now.)

However, I have a segmented run time of approximately 21:11:00.  (The time's not entirely accurate, but it should be accurate to within a few minutes.)

I'm new at speedrunning in general, though.  So I have a few basic questions...
1) Does anybody know if anyone has actually run this game before?  I couldn't find anything online.
2) If I die in a single-segment, is it acceptable to reset and reload as long as the timer's still running?  (Losing half of my gold and spending even more to revive my party would typically be catastrophic.)
3) I wasn't sure about the actual timing I should be using.  Does this seem reasonable?
      - Start is after the new file creation, right before the opening cinematic cutscene.  (After naming the hero.)
      - End is as soon as the final boss takes the last hit.  (Excludes death animation and the entire ending.)

Honestly, I'd love to be able to run this at a marathon, or something.  But I guess I should actually record a run first, lol.  Hopefully, I'll have a good single-segment fairly soon.
Thread title:  
Edit history:
BayouBillyBones: 2014-03-26 02:33:41 pm
BayouBillyBones: 2014-03-26 02:24:00 pm
Life's short. Get a dog.
For everything dragon quest, Nico Nico is the hub.  Here's the first run I saw, which may have since been obsoleted, it is a 13:49:34 japanese timing(starts from the powering on of the console, includes putting the ps2 to fast disc speed, and runs through credits):

Nico accounts are free, but if you don't have one, just add ".am" after the jp to watch the videos i.e.:

You will want to search "DQ7 RTA" on Nico for the best results.

I would say start by learning their route, then try to see if you can improve it in anyway.  DQ Nico runners in general don't like to use death abuse or glitches, so keep that in mind as you analyze the route.

Also, there is a 3ds remake of the game, but so far it is japanese exclusive.

EDIT: here are the notes from the run I posted:
Thanks, I'm going through it now.

13:49:34 seems beyond absurd, compared to the US timing.  I wouldn't be shocked if most of that's just the difference in the dialogue, because there's a ton in this game.  I guess I'll see.  I might just need to track down the Japanese version.

And I've been playing on a PS3 because the loading times in this game aren't too bad, but it's nice to know that I can use a PS2's fast disc speed.  I thought it would glitch this game.

And yeah, I haven't even seen the 3DS version.  I heard they changed a lot of the battle mechanics, though.
This will take a while to get through, but my initial impressions from the first two pedestal areas...

Overall, the strategy's really similar to mine, just a little more aggressive.  He's picking up fewer items than I did, and leveling at different times (but not less...), but there's considerable overlap.

It's clear that he's practiced this game quite a lot, though, because he makes very few mistakes.  It's also heavily optimized, even down to subtle camera changes and fast menus.

Glancing over the notes, he seems to grind for similar skills (like SwordDanc and Stampede, based on the number of battles), and even grinds for a few *more* that I hadn't been getting.  But he also apparently skips all of the TinyMedals, which is pretty gutsy... That skips a lot of great items, including the SageRock (free full-party healing).  And it'd be pretty hard to get through the end-game without some kind of party-healing skill.

But the dialogue is definitely much faster.  The shorter text alone seems to save about 4 minutes even before completing the first pedestal.  The disc speed also seems to have a subtle effect.

I made a fundamental error, though.  I was using Battle Speed #1 (fastest auto-scroll).  I should be using Battle Speed #8 (stops after every text box, button mash to proceed).  It's tiring, but makes battles considerably faster.

...And so much for no death abuse on Nico; he kills himself in the very first dungeon.  >.>
I want off the ride....
^ he said usually. But from what i've noticed they will die in the first dungeons in Dragon Quest runs because you can't really afford much and its a faster way back to the home base.

My best advice to that you should figure out what good calculated risks you can take. I'm sure everything he does is for specific reasons (even skipping the Tiny Medals). While you can't talk to him, it might be a good idea to try to figure out the reasoning for his route. Once that is done i'm sure you'll figure out ways to squeeze more time out of it. Its a bunch of give and take.
Ah, I figured out something.  He gives Gabo HealUs.  That explains skipping the TinyMedals/SageRock.  I had definitely considered using the Healer Heart to learn HealUs (160 battles), and Gabo has the agility to be a great healer.  But, I thought that Gabo wouldn't have enough MP for it to work.  I'll see how he deals with that.

I'm going through each part and taking notes on his strategies and decisions.  It's pretty hard to watch this game in Japanese, and I have to pause repeatedly just to catch his menu selections, so it's slow going.

I've noticed that the fast disc speed seems to speed up attack animations, as though the game loads each animation individually.  Not positive about that, yet.  It definitely speeds up the Return spell, but that's minor.

Also, he's using a Turbo controller.  That explains how he shoots through the battle text and certain menus so easily.  This game seriously requires a ton of button-mashing, but I guess Turbo's banned.  (But if there was ever a game to use it anyway...  Yeesh.)
I want off the ride....
Japanese community doesn't ban turbo on a controller. If you want to use it for a run for fun. Go ahead. Its if you want to submit you'd have to go "all natural". So if its easier to learn with it then ween off, do so.

I know your pain with the japanese, a suggestion might be to try to "play along" to your best ability. Thats how I had to learn the full route of Sweet Home from the J Tas (the menus were kinda hard to follow at times). It also gives you a source of how things should look and it'll help point out little differences if they are some (this item isn't in xxxx chest instead its at yyyy, so need to alternate here) and such.
I don't know this game well enough to comment on routing, but I'm excited to see that it has a serious runner!  I wish you good luck in finishing a run.
You just need to watch the speedrun and take the same steps with the US version, that's what I did with DQ8. I also managed to take a screencap of all the JP items names and rename thems kn with their US names  like this :

It helps a lot, you'll always know what item he's moving/selling/equipping this way just by looking it up, and since Dragon Quest is a pretty static game (aka unlikely to get random drops), you'll be able to go through the whole route this way with little hiccups.
I understand some fundamental Japanese, so it's not too hard.  Most of the item names are pretty simple, and the character names are direct translations.  I also remember most of the item locations and attack animations just from running it myself, and haven't noticed even one difference between the versions.  I'm about 10 hours through watching it, now.

I'm definitely learning some tricks, like a few extra uses for items, and a walk-through-walls glitch that I'd have never found on my own.  There isn't much accurate info available for this game (not in English, anyway), so it's pretty handy.

But, the biggest difference between his run and mine is that he picks up very few items in the second half of the game.  That's fine for most of the way, because the middle of the game is honestly pretty easy if you've been grinding for the right skills.  But, I'm very curious (and a little confused) as to how he'll deal with the final boss.
So, I finished watching it.  As it turns out, I completely underestimated the Wizard Ring.  It can be used multiple times, and restores 20-30 MP, so Gabo's low MP isn't really a concern.  (I think the Wizard Ring can break, but the game hands you several of them.)  Gabo can cast HealUs about 2 out of every 3 turns if he has to, so the later bosses suddenly look a lot easier.  (You still need to grind 160 battles to learn HealUs, but that doesn't take too long if you know what you're doing.)

Also, the final boss uses an attack called Glint, which puts one character to Sleep, completely bypassing Sleep resistance.  It's hard to cure Sleep in this game, so that's normally pretty obnoxious.  But, there's a weird bug.  Glint is apparently counted as a Paralysis skill (probably how it bypasses the Sleep check), so anything that can cure Paralysis can actually remove that Sleep status.  Specifically, NumbOff and MoonHerbs.  That makes one of the worst attacks in the entire game a total joke.

So, yeah.  The final boss still isn't exactly easy, but those two things definitely make him much more tolerable.

There's a lot of weird things I learned from watching this run, too.  Vertical movement is faster than horizontal movement, so you can turn the camera to walk faster.  (Probably about 30% faster.)  The Sorrow Shield's curse actually splits damage between party members, so it can be used defensively.  You can walk through walls in a single floor of Baloch's Tower, by pushing a boulder in a particular way.  And there are three separate places where killing yourself is faster than walking.  (One's obvious, but I'd have never thought of the other two.)  And, a few other more technical tricks.

Doing more research on this game, I also found out an interesting skip.  You can bypass the Probina boss battle.  Basically, there's a cutscene at the Probina Bridge, where the character Razuel breaks a statue.  If you talk to Razuel at the same moment you get in range of the cutscene trigger (might be frame-perfect, or pretty close), it triggers dialogue that isn't supposed to happen yet.  That, in turn, triggers Probina to be "saved" and appear in the present.  You still need to advance the plot until you get the Priest Key, but you get to skip the boss.  (You only get one shot at this skip, unfortunately.)

I tested the Baloch's Tower glitch, the Probina skip, and the Glint paralysis bug on the US version.  They all work.

The Japanese runner's strategy is pretty solid, especially the item management, so I think I'll try mimicking it once.
So, I've been running the game (segmented for now) using the Japanese runner's strats.  There's another huge difference that I forgot to mention: When he doesn't need to grind, he runs from almost every battle.  I thought that would be relatively minor because there's still a lot of required grinding, but I was very, very wrong.

Predictably enough, Verdham and especially Dharma are a huge problem.  Successfully beating Cavemon, Inopp & Gonz, the Arena, and the Dharma leader Antoria in one try and all in one run is practically suicide.  It's not impossible, but, still.  And, the Arena and Antoria are especially sadistic; I died about a dozen times there, and eventually just took a loss against Antoria (3-5 minutes and a lot of Gold lost).  And that's even allowing for the available 3 WorldLeaves and 1 Wizard Ring.

(I should note that even the Japanese run that I watched was not without flaws.  He even game over'd twice, just not while fighting bosses.  So it didn't cost him much time.)

My current time after beating Antoria is about 5:30, give or take 5 minutes.  (I ruined the timer resetting after all of those deaths.)  That's about an hour faster than my old time, though still over an hour slower than the Japanese runner's time.  (I think about 80% of that is just be the difference in Japanese dialogue and turbo, though.)  A lot of that speedup might just be the battle speed, faster movement, and menu optimization, so I'm strongly considering leveling a bit before the tougher parts.  The early levels come quickly and give really nice stat boosts, too.  But that's almost certainly a slower strat, so I really don't want to resort to that.

It should get easier for a while after Dharma, because I absolutely have to grind for skills.  That's essentially required for both speed and survival.  (I fear for the end-game, though.  I'd hate to run this game for 18 hours just to get stuck at the end.)

...Btw, I have a technical question about the PS2.  Would Texture Smoothing slow down the game at all?  I think it looks a lot better with it on, and I can't see any slowdown.
I want off the ride....
It probably won't slow down the game but you might wanna try to count how long loads take. Its something that seems really asinine but its a factor with a run as long as your going. Take a stopwatch or even (if you can with what you got) record a few sequences of loading and time them out. It might be better with using recorded footage and just use the same "sections" with smoothing on/off and fast disc on/off. And see if it makes a difference.

I say this because not every ps2/game is the same. Some games do worse with those options enabled, some do better. Also different PS2s can be awkward (some slims are better or worse depending on the game as well). So just perform a test if you can with the video.

As a suggestion, try to see if you can cope with a few more battles and gradually wind down. Sometimes its just a matter of getting used to being in a tougher spot. Wind down on the grind, don't bypass it completely! you never know if things might be better with that +1 level if it didn't take too much.
Thanks for the info.  This game is weird for a disc game, though, because it doesn't seem to have any major loading sequences.  (Except maybe the portals, which are long sequences, but always seem to take the same amount of time.)  Instead, it seems to load each area and animation on the spot.  Or, it might cache them as you proceed.  (The only time I've ever noticed the game *stop* to load something is if I leave it on in the middle of a battle, and stop playing for a long time.  Then, the disc skips and the game stops for about 3 seconds, while it loads the battle animations.)

That makes me suspect that the disc speed could save small amounts of time for each area and battle animation, but it's a little hard to measure, especially in battles.  I'll run some experiments eventually, though.

...Anyway, analyzing the strategy a bit further, there's a weird gap in the experience levels just before Dharma.  Particularly for Gabo, who's the weakest link because he joins late.  He starts at Lv1, reaches Lv4 really quickly, takes forever to hit Lv5 and Lv6, and then levels fairly quickly again until about Level 12.  It's hard to explain, but basically, leveling doesn't help as much as I thought.

Instead, as an experiment, I changed back to my old items and more offensive Arena strategy, with the same low levels... and won rather easily.  In particular, the Japanese guy wastes a lot of turns trying to land both Sleep and SpiderWeb/Bark (enemy loses a turn) on Antoria.  Landing both is nice, but SpiderWeb/Bark usually won't work on him.  So, I just put Antoria to sleep repeatedly, and kill him outright.

I might also sell the first Wizard Ring (1500 Gold, which is quite a lot) instead of using it as an item in the Arena.  I can usually get through the Arena and Antoria without restoring my MP at all, and there are some really nice items for sale in Dharma, so I think it's worth it.  There are 5 free Wizard Rings in the game, and 4 are practically handed to you, so selling one isn't a big deal in the long run.  (Actually, 1 of those 4 is *literally* handed to you, lol.)

I'll keep trying different things, though.  Dharma is definitely one of the biggest hurdles in the speedrun, so a good strat there is important.
Edit history:
Emily: 2014-04-03 09:39:08 pm
I've been testing the PS2's disc options.  I have the original PS2, not the slim.  I'm not really sure what the differences are, though.

Smooth Texture Mapping appears to have absolutely no effect on the loading speed.  But, the disc speed is giving me some pretty strange results...

I did 3 tests, several times on the various settings.

1) 5 times in and out of a small town.  I'm consistently getting about 1 second faster on the Fast disc speed.

2) 5 Return spells to the same location.  This is harder to time properly, because I have to go through several menus.  But the Fast disc speed appears to be 2-3 seconds faster.

3) 5 times in and out of a portal.  This one's... weird.  I'm consistently getting times 5 seconds SLOWER for the Fast disc speed.  In other words, I'm losing 1 second per round trip.  I tried it 4 times on each setting, and using 2 different portals (into the smallest area, and into one of the largest), and kept getting the same exact results.

I don't really understand these results.  (I have noticed, though, that I can hear the disc skip on the Fast speed while entering a portal.  That's probably related.)  Over the course of the run, I think I use portals (round-trip) 25 times.  That's apparently about 25 seconds lost.  But, I think I'd lose even more time from all of other, normal loading zones.

25 seconds is pretty minor in such a long run, but I'm wondering if there are other places that it's slower.

I didn't experiment with in-battle animation loading times.  I couldn't think of any way to test that.
I decided to keep going after Dharma.  I'll work out the details of the strats later.

And that means... skill grinding!  I fought 205 battles in just over an hour.  (I don't have an accurate time, because I also had to clean out and leave Dharma, and visit Mezar.)

...That's actually a really good pace, now that I think about it...  Over 3 battles per minute.  (You can hit that pace in Engow or Dialac, but you must be at a low level for it to work.  Maximum of Level 13 for Engow, and Level 15 for Dialac.)

On Battle Speed 8, it's basically an hour of straight-up button-mashing, though, so my hands hurt a bit now.  My time for the grinding was about 2-3 minutes slower than the Japanese runner's time, even though he was using a turbo controller.  (You could easily just mash X in battle, so it'd be one of the best places to have turbo.)  That's a pretty good pace, I think.

Following this strat, I'll have to fight 65 more battles for skills later on.  But that'll be pretty easy.

For those wondering...
53 battles: Hero - Dancer, Gabo - Thief, Maribel - Shepherd
82 battles: Hero - Warrior, Gabo - Healer, Maribel - Shepherd
13 battles: Hero - Warrior, Gabo - Healer, Maribel - Mage
57 battles: Hero - Shepherd, Gabo - Thief, Maribel - Mage

30 battles: Hero - Warrior, Gabo - Healer, Melvin - Mariner, Aira - Dancer
35 battles: Hero - Warrior, Gabo - Healer, Melvin - Mariner, Aira - Warrior

Honestly, I had a really hard time finding a good class path before, but this path seems incredibly solid.
I want off the ride....
Quote from Emily:
I didn't experiment with in-battle animation loading times.  I couldn't think of any way to test that.

Do you have an endgame file that you played for EVER with? A suggestion might be to go to a place where enemies use techs/skills often, and have each guy spam their own spell. Run it about 20-25 times maybe? to account for different actions on the enemy side (or better where you will get to use all of them and kill the enemy before they act). Then time it out. If you are using the same techs each time it should take the same amount of time either way. Might be better to use a turbo to get best effect out of this.
I forgot a few portals in my precious count, because you have to re-enter some to grind.  It's more like 30-35.

Anyway, I did some battle tests.

Quote from RaneofSOTN:
Do you have an endgame file that you played for EVER with? A suggestion might be to go to a place where enemies use techs/skills often, and have each guy spam their own spell. Run it about 20-25 times maybe? to account for different actions on the enemy side (or better where you will get to use all of them and kill the enemy before they act). Then time it out. If you are using the same techs each time it should take the same amount of time either way. Might be better to use a turbo to get best effect out of this.

Yeah... I have a file like that, but it's on my PS3.  I think there's too much randomness in enemy animations anyway, but this gave me a good idea.  A test where your party just uses the same attack to end a battle over and over would give consistent results.  I don't have many attacks like that on a low-level run.  But, Time Sand instantly resets the battle, which is kind of the same thing.

I also realized that entering and exiting the battle could add some loading time, so I tested that, too.  So, 2 more tests...  And, I used Battle Speed 1 (auto-scroll) to eliminate my own button-mashing.

1.) Whistle and run from 5 battles in Rexwood (the first area), to test the loading in and out of the battle itself.  (Running from really weak enemies is guaranteed, so it works.)  5 battles seems to be about 3.5 seconds faster on the Fast disc speed.

2) Use Time Sand with the fastest character, to reset the battle 5 times in a row.  There are more factors here, namely the menu time, so I couldn't get a very accurate time.  The Fast disc speed appears to be a tiny bit faster, maybe up to a second.  (Or, they might be roughly the same; I'm not really sure.)  It's not a very significant amount, and it's hard to measure.  (And naturally, the loading time might depend on the attack animation itself.)

Regardless, that battle loading speed is very significant.  I think I fight (or flee) roughly 500 battles over the course of the run.  (Give or take a hundred or two...)  That's somewhere around 6 minutes just for loading in and out of battles.  So the Fast disc speed seems to be much better.
I'm still working on the practice run, based on the Japanese runner's strats.  I'm over 2/3 of the way through (entering Coastal at about 12 hours; around 2 hours slower than the Japanese runner's time).  Dharma was a nightmare, but the middle of the game is really quite easy with this class path.

I think I'll get a final time of around 16:15:00-16:30:00.  That'll depend on how Disc 2 goes.  Granted, this is segmented for now, but I think a 17-hour single-segment is looking quite plausible.  Under 16 hours might even be possible, but only with particularly fantastic RNG.

It's much better than before, either way!  The faster walking, and skipping so many items, seems to have sped things up tremendously.  I plan to attempt a single-segment soon after I finish playing through this run, but it might take me some time to setup.
I finally found a difference between the Japanese and US versions.  The game-ending Rainbow Dew bug has been removed.  And, the way that it's been removed has changed the speedrun strats slightly.

In the original Japanese version, when a character leaves your party temporarily, they take some of their inventory with them.  The game removes certain items, including important key items, and places them in your Bag.  Unfortunately, the designers forgot to move the Rainbow Dew, which can be obtained very early in the game, but isn't used until close to the end.  If Maribel takes the Rainbow Dew with her when she leaves the party, you'll be completely stuck when you reach Coastal.  (I've never seen this bug personally, but I've seen it mentioned elsewhere.)

In the US version, this bug has been fixed.  ...Aggressively fixed.  Instead of moving only certain items, the game takes *all* of the character's unequipped items, and places them all in your Bag.  Your characters are frequently shuffled in and out of your party on Disc 2, so this completely changes the late-game item management.  It also gives you fewer WorldLeaves to work with.  (You can get a free WorldLeaf if you're not currently holding one, so Japanese players can trick the game into giving them a few bonus WorldLeaves.)

It's a minor difference, but I just thought I'd mention it.  (It appears to be the only difference, aside from the dialogue.)
I want off the ride....
Quote from Emily:
It's a minor difference, but I just thought I'd mention it.  (It appears to be the only difference, aside from the dialogue.)

Nothing is quite that minor. But its a difference. I'm sorry for not having played DW7; i'm just cheering ya on.

Remember even little details add up when going for something amazing. The smallest of cracks can tear down a wall. Good luck and keep it up!
Final time: 15 hours, 50 minutes!!!

Final time including the ending and credits: 16 hours, 11 minutes.

Much faster than I expected!  Remember that this was a segmented run, though, and I reset several times for better RNG.  Ultimately, it was done in 16 segments.

Like I said, the game is really easy for a while, once you get past Dharma.  But, the difficulty jumps again at the end of Disc 1 (about 3 hours before the end of the game).  Several late-game bosses (Nengal, and both of the Orgodemir [demon lord] battles) are extremely dangerous, and you're severely under-leveled by that point.

To deal with that, you want to kill at least 1 Metabble in Coastal on Disc 1, and probably about 3 MetalKings in the Dark Palace at the end of Disc 2.  (You can get by with 1-2 MetalKings, but the final boss is pretty freaking nasty.)  But, anyone who's ever played a DW game knows how finicky metal enemies can be.

It's unfortunate that the end-game strat relies so heavily on the RNG, but MetalKings give a ridiculous amount of experience.  I got lucky and killed 4 MetalKings in my segmented run, and every party member gained 9-10 levels, going all the way from Lv20 up to Lv30, in just a few minutes.

Orgodemir himself has some very distinct attack patterns, but most of his attacks are still quite random.  Ultimately, he's just extremely powerful and takes a long time to kill, so you just have to defend as best as possible, and heal away the damage and status effects as they come.  As for Nengal, he's just a very strong and very random physical boss.  He uses the always-broken SwordDanc and randomly lands critical hits.  But, you can survive that battle more easily by having 1-2 people parrying at all times.

I still don't have any means of recording a run, but my father has agreed to help me with that next weekend.  I'll figure something out, either way.  I won't attempt a single-segment unless I can record it, but I'll keep fine-tuning the battle strategies in the meantime.

Quote from RaneofSOTN:
I'm sorry for not having played DW7; i'm just cheering ya on.

Thanks for the support!  It seems like a lot of people haven't played most of the DW games, which is kind of strange, given how popular it is in Japan.  DW7 isn't the best game in the series, not by a mile, but it is one of the harder ones.  I also thought it would be an interesting novelty to speedrun, because it's just so incredibly long.  Most people take 100-200 hours to beat this game, and I think it might be the longest game ever made.  (Well, the longest game with an actual ending, anyway.)
I'd also be interested in seeing this done (as a huge fan of the series and 7 in particular!), and I'm impressed that you're intending to single-segment such a game. I'm surprised that it can go as low as 15:50 in English, due to the sheer amount of text in this game. I always got Gabo to TeenIdol for Hustle, but you're not putting in anywhere near enough battles for that. What is the advantage in your class path, out of curiousity?
Quote from MaskedHuzzah:
I'd also be interested in seeing this done (as a huge fan of the series and 7 in particular!), and I'm impressed that you're intending to single-segment such a game. I'm surprised that it can go as low as 15:50 in English, due to the sheer amount of text in this game. I always got Gabo to TeenIdol for Hustle, but you're not putting in anywhere near enough battles for that. What is the advantage in your class path, out of curiousity?

Hustle would be so very wonderful to have, but that takes 455 battles, and as such isn't really an option.

My current class path (Grinds 270 battles) teaches the following useful skills:
Hero - SwordDanc, EvilSlash, Whistle, Mineuchi
Gabo - Heal/HealMore/HealAll/HealUs, Tiptoe, Knockdown
Maribel - Stampede, Blazemore, SleepAll, Whistle
Melvin - NumbOff
Aira - SwordDanc

SwordDanc and Stampede are magnificent attacks, and Blazemore is good when Stampede won't work.

HealUs is needed for healing.  Attempting to finish the game without some sort of party healing skill is just suicide.

Knockdown removes an enemy from battle, but actually checks for Death resistance, not Expel resistance.  It also has 100% base accuracy, although that's lowered by the resistance factor.  Certain weaker required enemies can be completed eliminated by Knockdown.

Mineuchi and NumbOff are status-healing skills, and can be very helpful against the final boss.  As stated previously, NumbOff also removes the effects of Orgodemir's Glint.  (NumbOff actually requires 10 more battles not listed.  You can get that easily while Metal hunting.)

Tiptoe reduces the rate of enemy encounters.  However, it also increases the rate of surprise attacks, so it's a double-edged sword.  Whistle is naturally very useful for grinding, and EvilSlash is just for killing Metal enemies.

Aside from the class skills, the characters (especially Melvin) have some pretty useful natural skills.  Gabo doesn't learn any good damaging skills from the path, so instead he uses Howl and the Flame Claw (casts Blazemore when used as an item).

The final party, by the way, is Hero/Gabo/Maribel/Melvin.  Aira is left behind.
If SwordDanc works on Metals, consider trying it. I never tested it as hard in 7, but in 9 skills that multihit and can crit can take out Metal family no problem. It's been years since I last played 7, though, so I could easily be wrong in that regard.