Posts for oneeighthundred


Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
The first one that really got me was that "glitchfest" playaround of SMAS SMB3. It was constantly doing unexpected, creative stuff and kept it up for a solid hour. Runners up would probably be the SM RBO runs (which accomplish kind of the same thing) and the first Chrono Trigger with subframe reset (which was a small miracle considering how many bugs had to be used to win or escape every mandatory battle).
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
Any game that emphasizes reaction time and precision has a lot of opportunity to be torn up by a TAS without too much down time, so most action games are good choices, especially ones with projectile attacks and speed boosts. I'd actually disagree about going with games with known tricks, new games are less competitive and give you the opportunity to make something new and interesting even if all you know how to do is route planning and being super accurate. There are several threads in the games forums with ideas/wishlists of games that haven't been done, and there's a wiki page for Genesis with some good choices bolded.
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
Warp wrote:
Can someone give an actual argument why the category should be removed, other than "it's stupid", which is just childish?
It doesn't say much about the run (especially since most of what AVGN complains about in games are completely avoided by TASes), and AVGN isn't sending traffic here, so it's just a name drop. Since it's a name drop, whether it's good/bad is going to depend entirely on whether people like or don't like AVGN, and not everybody likes him. (I don't really care, personally, but I can see why people who don't like AVGN would be bothered by it)
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
adelikat wrote:
So how about this: Gold medal = old star criteria Silver medal = old moons if they were used with consistency Bronze medal = current moons Vault = current vault
Well, the "milestone" moons weren't really on a scale with the others (as gold/silver/bronze would suggest). They should probably be a separate thing entirely, like A2600 ET might be notable for being the first A2600 run even though it's vaulted, while discovering a new major glitch might be notable and yield an extremely entertaining run.
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
adelikat wrote:
So this movie is an interesting case for the the Vault. The logic for abusing death in cases where there is infinite continues from the place you left off is sound, for entertainment value. But the vault isn't about entertainment value, so should this rule not apply? A deathless record is still the most impressive record here is it not? We should strive for the vault to have impressive records, even if entertainment value is not considered.
I don't think deathless should prevent a Metal Slug game from being published, but quarters are a tricky case. On one hand, it's somewhat arbitrary and the console/handheld versions give you infinite continues. On the other, it's based on an arcade game where quarters are a non-gameplay way of stopping a Game Over, and the game is totally impossible to lose with continues, which seems to go against the entire point of difficulty requirements if not the exact wording of it.
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
I'm trying to handle the encodes for this as well (and maybe do more in the future), but I'm not entirely sure what I'm doing, so if someone could tell me if/how I'm screwing this up, it'd help. MP4s: https://www.dropbox.com/sh/blnsyfik5p85ulw/nxDc93JJI8 HD YouTube's in the sub.
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
adelikat wrote:
jlun2 wrote:
adelikat wrote:
No because it wouldn't be a vault submission, it would be _a_ submission. Your vote would be "I did not find this movie entertaining, no". And without sufficient audience response, but still a speedrun record, it would be a vault publication.
Can publications obsolete a higher tier one? Or would it be separate?
I don't see why not.
A possible reason not would be that a vault-tier publication obsoletes a moon/star-tier run that has a goal choice that's disallowed in vault tier. How would that play out?
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
I'm sort of concerned with the uniformity of the moon tier obsolescence criteria, mainly with how speed/entertainment tradeoffs are concerned. There have been runs like the subframe reset Chrono Trigger that were much less entertaining because of what made them faster and there's no reason to think it won't happen just because a goal is "good." Speedruns by their nature improve by skipping potentially-entertaining content. IMO, new goals should be approved if a moon-tier submission can be done, but once approved, should not prevent vault-quality runs from obsoleting them unless the goal is revised to bar something that made the run less entertaining (and the bar should have something to do with the goal).
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
I'm back (again). Round 2 is clear, now with 50% more useful character variety. It turns out that the mages and slimes in 2-7, 2-8, and 2-9 can all be hit by multiple attacks in the same frame, which saved some time. Have a video: Link to video
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
The FPS portions of Nightfire used a modified goldsrc (a.k.a. Half-Life 1), which is a Quake-lineage engine. I know you said that jumping didn't provide a speed boost, but did you see if speed accumulation through strafe-jumping worked in this title? It might not be possible with the GC controller sensitivity, but if it is, it would probably save a lot of time.
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
I'm not dead, just distractable. I'll be uploading some more progress this weekend. e: Updated! One major change: Mars' death in 1-10 has been undone and his death has been moved to 2-5. Fortunately, this game seems to be extremely resilient to editing, even touchy stages like 2-3 are staying synced with edited input. It's not clear if Fire Knight will be useful (the next few stages will probably determine it), but neither of the Mars picks cost any frames on the current route so I'll probably leave him dead.
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
OK I'll go through what worked for me: - Start the game up, get Mario to a dead stop - Pause emulation - Save in slot 1 - Hold down the Right button, advance until Mario moves 1 frame to the right, then 1 frame further so you know you get some subpixel movement in but not a pixel move. This wound up being 6 frames. - Save in slot 2 - Load slot 1 - Advance the same number of frames (6) - Open RAM Search, Reset, Clear Change Count - Load slot 2 - Set filter to Greater Than, Number of Changes, 0 - Click Search For me, this knocked the RAM value list down to just 25 values. Unpause emulation, wait until Mario comes to a stop, then watch the remaining values and see what continues to change. If something keeps going up in change count, it isn't movement-related. Eliminate it. Pace around a bit, some values will rapidly increase in change count. Some values only changed while changing direction, some changed a lot. I got a few of them up into the hundreds, then eliminated anything below 100. This got it down to 10 values. Of what's left: 7E005D: Goes down to -24 when full speed walk left, increases to 24 when full speed right. This is obviously your X speed. 7E0219: Changes entirely based on X location on screen. This is your X coordinate, or at least related to or part of it. 7E0290: Changes between a very small number of values in a short range, which isn't typical of subpixel behavior. 7E0700: Always the absolute value of 7E005D. This is your abs speed. 2 of the values are highly volatile, i.e. they cover a very wide range of values: 7E0401 and 7E0705. Of those two, 7E0219 changes value every time 7E0401 crosses 0. This means it is probably your subpixel X. Y is harder to tell since it changes so much from gravity, but it's probably either 7E041C or 7E043C.
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
I'm not sure if this applies but if SNES9x has Gens' ram watch stuff then it should I guess. Easiest way I've found to dig out values is load a state, do something known to change a value you want to find (i.e. just walk left for a couple frames), save the state, reload the earlier state, don't do that thing for the same number of frames, save that in another slot, reset your change count in the RAM watch, load the other state, and then drop everything with zero changes. This gives you only values that are different between the states. This should severely reduce the number of values listed. Then, don't do anything that would change the value for a while (i.e. stand still), then drop anything which still increased in change count, since those are changing even though you're not moving. Watch the remaining values (there should be very little left), try to figure out how they relate to your movement, and you should find it.
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
Orange Claw Hammer wrote:
looks great, glad to see some nice progress on this!!! btw, blechy and i (hi im cyghfer) are considering doing a 2-player console run of this game in the future, and a completed TAS of this would help us immensely in planning such a run. seems like your motivation is pretty strong but any extra couldn't hurt!
The non-TAS strategy is going to be pretty different from the human run strategy since you'll be suffering deaths and won't be able to dodge every projectile. I can tell you roughly what you'll need to know though: - Read the earlier posts in this thread about HP levels and how much damage attacks do. - Do not lose your archers. They have the highest speed, the best basic attack range, there are enemies in rounds 2 and 4 that can only be hit by SAMs and arrow specials, and there are enemies in round 4 over barriers that can be hit by arrow specials and nothing else. If one of them gets hit, your next pick should be a SAFE choice to get them to the exit. Don't over-pick them either, the riflemen and machine gunners have similar range but are less valuable. - Diagonal spears only have a 2px range where they can hit up most ledges, mere mortals should stick to verticals or use an archer because otherwise you'll probably miss! - Robby's missile altitude depends on stage! If the stage has enemies on ledges, it will be high altitude and ONLY hit those enemies. If not, it will be medium altitude, colliding with most barriers (but not low barriers like sandbags) and hitting most enemies. - Don't pick a slow character that will be under projectile fire or rushed by enemies. They can only have 1 basic attack out at a time, are slow, and their specials are bad at dealing with close enemies. The best characters for dealing with rushing enemies are magic knights and machine gunners (due to their high basic attack limit), the best characters for dealing with projectile fire are riflemen, machine gunners, and archers (due to range). - Don't pick the flamethrower guy ever. - Magic knights are really good in round 2 because there are a lot fewer projectiles and a lot more bags of health. They have BAD range though, so they're not very useful in later rounds. - Water Knight's stun normally lasts 3 seconds, but resets to 12 frames (0.2 sec) when an enemy's damage immunity phase ends. If you're trying to use it to gain distance from a fast enemy rushing you (i.e. orbs, knights, slimes), don't attack them until you're happy with your range. - You can one-shot the 2-10 boss with a tornado launched from the center, which is fast, but a bit risky since it doesn't always work and he'll be on your ass if it doesn't kill him. Bonus points for having a magic knight to kill the knights on the lower level, but make sure your other player can hit the ledges. - Make sure you pick someone with a high-altitude attack on 3-10. Mud Puppy (the guy with the north-only long-range special that does triple damage), is great for killing the bosses but can't hit the ledges. - Cyber and Mud Puppy are really good in MOST of round 5, but bad at dealing with enemies behind barriers. - Most ambushes have an aggressive AI type, they will rush you, so you should prioritize killing ambushes over killing stationary enemies. - Grid-fire projectiles in round 5 can be intimidating but really, figure out where you need to stand between two rows of vertical fire, figure out when you can move further up without getting hit by diagonal fire, and take your shot. Alternately, pick Cyber or Mud Puppy and just kill shit from out of range. - Read the earlier posts for how to dodge the last boss's missile attack.
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
Game: Gain Ground System: Genesis Author: oneeighthundred Completed: Round 1 (of 5) Discuss: Here WIP: Link to video
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
The phone company has decided to incapacitate my DSL for a full week. Fortunately, I have other computers I can use to post my WIP. All of round 1 is clear in roughly 2:27. OP updated with a new GMV. 1-7 and onward use a new orders bot that should make it easier to rebase things if I wind up having to change earlier levels and make fine-tuning the weird immunity frames easier. I upgraded it with spin functionality in 1-10. I'll probably be updating the video WIP every round (the GMV is still every stage, pending my DSL getting fixed of course) Link to video Also, in case anyone's wondering why Athra/Gascon are not just killed immediately when Verbal can kill enemies faster: First, Gascon doesn't have a diagonal spear special, which means he can't clear ledges or kill the demon turrets as quickly, so killing Athra without killing Gascon slows things down. Second, spear specials take less time (~9 frames) to land than arrows and 1-6 is constrained by special flight time, something that rarely happens later in the game. Third, most stages take movement to even get them to a spot they can even be killed at, adding to the already-high cost of killing them when a stage is about to end.
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
6 stages done, OP updated. 1-6 was an interesting ride. Made a HUD script: http://icculus.org/~riot/gainground.lua Puts a 16x16 box around everything (correct collision size for everything except bosses), displays AI type and timers, and shows projectile collision locations. Slightly glitchy due to the inconsistent memory layout of enemies. All projectiles are 1px, including fireballs and bugs, so if you can avoid the death pixel you're golden. edit: Okay, projectile display is only accurate for projectiles traveling horizontally or vertically at the moment.
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
arkiandruski wrote:
I kind of wonder whether hard or normal would be better. My argument for normal is that you don't start with everyone on the roster, which means that you have to pick people up if you want them, which means strategy will change over the course of the run, and you have to make decisions on whether the time taken to get a person is worth the time he saves. That seems much more interesting to me. Also, there may be levels on normal where just going to the exit is faster. Not sure if that last point is a plus or a minus.
Yeah it's kind of a tough call. I don't think it'll make that big of a difference in practice: There aren't really that many useful characters and Medium gives the two that I want the most (a magic knight and Verbal) in the first two stages. The only other character that seems really important is Cyber in 2-9, and maybe Fire Knight for certain bosses. I went with Hard because I think character combo variety will make things more interesting in early stages, there's still strategy consideration in that Hard's roster takes time to scroll through so l'm looking for opportunities to lose characters. Medium wouldn't have that since the player not running a captive has more time to get killed, and there are fewer characters that need to be (mainly Athra and Betty).
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
Doing a quick run through the game to figure out roster optimization, enemy HP levels, phasing behavior, and levels which can viably be used to kill off characters. It looks like this will be MUCH more important than previously thought due to newly-discovered damage variations, and with some stages having walls that can only be passed by archer specials and Robby missiles. Damage: - Most attacks do 1HP damage - Boomerangs, grenades, knight basic attacks, Fire Knight's special, missiles, and grenades do 2HP damage - Mud Puppy's laser cannon and Glow Knight's shield orbs do 3HP damage - Water Knight's special does 2HP damage to enemies that are immune to the stun effect - Tornados do 2HP every frame that they are in contact with an unphased enemy. The conditions for collision are unclear, but differ from the damage check. Large enemies can often be killed by a single tornado due to it not bouncing, especially the 2-10 boss. Phasing: Most enemies become "phased" when damaged for 24 frames. Their AI timer counts up USUALLY every other frame and the phase ends when it reaches 12. Sometimes it will count up on odd frames or several frames in a row, for reasons unknown. The same timer is used for the water knight stun, which counts DOWN, so hitting an enemy with a water knight stun and then damaging it will give you 12 frames to attack it again before the stun breaks. The enemy's state counter does not progress while phased, but does appear to progress on the frame it recovers even if it would be damaged. Orbs and spiderbots in round 5 have a 14-frame phase instead. Multi-hitting: Most enemies can be hit by multiple attacks in the same frame, but not all combinations work. Arc shots like arrows and basic attacks do not appear to multi-hit with basic attacks, but multiple basic attacks and Cyber rockets do. Some enemies (mostly bosses) are immune to multi-hit, only the first attack that hits them in a frame will register. Boss immunity frames are incredibly fucked up, delays for registering the next hit on a boss can be anywhere from 1 to 10 frames and sometimes it's immune to one player but not the other. I have no idea how this works yet. Teleporting: The next location that an enemy will teleport appears to always be in a set order. When an enemy teleports, it will stay hidden until a player enters its area of interest. Their AI timer is also set to a value that appears random, which counts down to zero (displayed as "Hide" in the HUD), during which it will ignore the player. Enemy behavior: Enemy areas of interest are always relative to the enemy. They vary considerably by level and enemy type, but appear to always be the same in a level for a particular enemy type. Check the AI type tag in the HUD to see what the enemy behavior is. Stage notes: Round 1: Most enemies have 1HP 1-9: Turret has 20HP 1-10: Turrets have 10HP, boss has 30HP Round 2: Archers have 1HP Warriors have 2HP Moat monsters have 2HP Knights have 4HP Wizards have 2HP Slimes have 3HP and can only be damaged while moving 2-2: Archers at the center have 2HP, door has 5HP 2-3: Probably an opportunity to lose a character. Only archers and SAMs can hit the rooftop enemies. 2-4: Wanderers have 1HP 2-5: Warrior ambushes have 1HP 2-7: Only arrows can go over the walls. Warriors have 1HP. 2-9: Walls do not block any projectiles 2-10: Boss has 30HP and multi-hit immunity. Round 3: Burrowers: 3HP Everything else: 1HP 3-5: All enemies can be hit by spears 3-7: Asymmetrical 3-10: Bosses have 30HP each and multi-hit immunity. Round 4: All enemies have 1HP Boss: 40HP 4-1: Asymmetrical, most ambushes are right-side. 4-4: Only arrows can shoot over the walls, but doesn't seem to matter since all enemies can be shot from the doorway. 4-5: Only arrows and SAMs can reach the top level 4-6: P2 must projectile kite due to an obstacle. Spears can hit all enemies, but Verbal can hit enemies on the opposite side. 4-7: Only arrows can hit over the walls, the trigger position for the ambushes is high enough to kill the enemies with basic attacks. 4-8: Only arrows can hit over the walls. Possible opportunity to suicide Mars on the top-left enemy. It is not possible to hit the row next to the exit from the bottom, but it is from the top. 4-9: Only arrows can hit over the walls. 4-10: Boss has 40HP and does NOT have multi-hit immunity. Round 5: Mechs: 2HP in 5-1 and 5-2, 3HP in all other stages Orbs: 3HP, 14-frame phase Spiderbots: 3HP, 14-frame phase Most crossfire grids can be shot through with Verbal by standing between vertical lines and waiting for a diagonal line to fade, then stepping forward. Orbs have a 14-frame phase and can be hit by multiple attacks at once. 5-2: Vehicle has 30HP and does NOT have multi-hit immunity. The walls do not block projectiles. 5-6: Mechs have a short area of interest. 5-7: Force field pylons have 6HP. Damaging them to 0HP disables the force field for 240 frames. It is possible to hit the second row of mechs from behind the lower force field with Verbal. 5-8: Laser turrets have 10HP each. Valkyrie can kill these in 14 frames each. 5-9: Gold orb has 9HP. 5-10: Barriers have 3HP. Boss has 30HP. The boss's missiles can be avoided by picking a max-speed character around when the missiles are at the middle barrier and going straight up. This will cause the missile to loop straight up, staying on screen and preventing a refire before you get inside of their loop.
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
Tornado will be pretty good if I can make it work, but it's REALLY niche and hard to compete with Verbal just putting arrows exactly where they're needed. Added a WIP link. 2 stages cleared. I think 1-1 is unimprovable, the top-left enemy dies last despite P1 making a beeline for it. Fixed 1-2's route to kill the last enemy sooner, saved 4 frames. Verbal must take the C shot at the frame he does, a frame later and the top-right enemy idles instead of moving down, taking longer to kill, a frame earlier and he misses. The main culprit for optimization and screw-ups is probably the no-move frames: Any direction change (including direction changes to fire projectiles) resets subpixel movement, causing a frame of no movement and potentially causing some later frames that would have been moves to be no-moves (since the slowest part of the move period is always the start).
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
I've always been surprised that nobody's done this game. It's a pretty nifty title that got several remakes and showed up in a bunch of classic packs, and is the exact kind of game where anybody running it is probably thinking "if only I had superhuman reflexes to dodge all of these projectiles and get shots in during 3-frame margins of error." I think the barrier to entry for a TAS of this is low, so I'll make this my first! Wish me luck! For the unfamiliar, Gain Ground is a game about killing stone age barbarians with missile launchers and robots with spears. Actually there is a plot but it only explains one half of that. The Genesis version has 10 more levels than the arcade version. I'm running Gain Ground (UE) [c][!] as NTSC, but this syncs with the Japanese version as well so whatever. The difficulty settings are weird, especially in 2P. Hard gives you all characters at the start but doesn't let you recruit any, which is "hard" because on lower settings, some characters can be gained more than once, giving you more chances in the later stages. I think Hard also removes a captive from the map when you die even if you aren't escorting it. Really, this means Medium and Hard are fundamentally different, some characters are unusable early on Medium and some can't be doubled, but character selection takes less time. A 100% run is also completely different since 100% on Hard just means kill all enemies, but 100% on Medium would mean kill all enemies AFTER rescuing. I'm just doing a vanilla speed run: Hard difficulty, 2 players. All stages will be cleared via kills, which is more of a guess than a goal: There are other strategies that might be faster, mainly suiciding everyone but one character and going straight for exits where possible. The problem is that it's a totally-different strategy and requires spending well over a minute killing characters. Given the current pace of the run, it's highly unlikely that a mass suicide is faster on Hard. I'm going to do 2P because it's faster and will probably be more technically interesting. Damage: - Most attacks do 1HP damage - Boomerangs, grenades, missiles, fireballs, and Knight basic attacks all do 2HP damage - Tornados do 2HP damage every AI frame (which is every 2 frames) that they are in contact with an unphased enemy. Some enemies do not bounce the tornado when collided with. Many (mostly bosses) do not collide or phase, and take massive damage. - Water Knight special does 2HP damage to enemies that it fails to stun - Glow Knight and Mud Puppy special attacks do 3HP damage Character walk speeds: Cardinal/Diag (in subpixels/frame) Spears/Archers: 192/134 Boomerangs/Grenadiers/Knights/Rifles/MGs: 144/100 Heavies: 96/67 Some thoughts/notes: - Explosives do NOT do splash damage - Scrolling past characters takes 25 frames - Killing a player takes about 146 frames - Most enemies will be killed instantly if you contact them during the character select invincibility. This doesn't apply to the demon turrets. - Character selection has a weird behavior where the ability to SELECT a character will be available 3 frames later for one of the two players, and this is independent of lag. (There is lag when selecting a character). However, scrolling characters is NOT delayed, so I'm generally trying to make the character eating the selection delay the one that has to change characters, where applicable. Which player gets the laggy select seems to vary. - If a player's projectiles are live when a player dies, character selection is delayed until the projectile disappears. - Death delays stage completion until the dead player can select a character. - Loose projectiles do NOT delay stage completion - C attacks can not refire until the projectile disappears. Fade-outs (i.e. tornado) and explosions also delay refire. - Handedness only matters when facing directly up or down. - Projectile are offset from your sprite. Characters with high B fire offsets (i.e. spears, archers) need 4px of clearance from an obstacle directly above them or they will hit it. - Projectile refires can be ordered on the frame that the sprite disappears. Watches: FFA801: P1 X FFA81F: P1 subpixel X FFA803: P1 Y FFA821: P1 subpixel Y FFA901: P2 X FFA91F: P2 subpixel X FFA903: P2 Y FFA921: P2 subpixel Y Subpixels are reset to zero on direction change (this also prevents movement for a frame). Stage change subpixel optimizations are not possible because character selection resets your movement state to stopped. I'm updating the WIP GMV every stage! You can get the current GMV here. 1-1: Athra+Athra 1-2: Athra+Verbal. Moving Verbal down at the end (on that exact frame no less) causes the top-right enemy to move down, allowing it to be killed 4 frames sooner. 1-3: Athra+Gascon. Gascon commits suicide to reduce selection time on Verbal. It looks possible to manipulate the top spear thrower to move left and be killed by a vertical spear shot before Ashra can contact it, allowing for 2 frames saved by better movement, but I haven't been able to duplicate that with Gascon selected. If Ashra is another pixel to the right, he'd get hit by the spear, another left and the bottom enemy wouldn't die from invincibility contact. 1-4: Athra+Athra: Needed to kill a few enemies from invincibility contact to prevent projectile firing from causing lag. P2 kills a ground archer early to prevent it from firing right and interfering with the ledge kills. 1-5: Athra+Athra: Not much to say. 1-6: Athra+Athra: And up goes my rerecord count! Tweaked P2's shot timing to avoid lag at the start. 1-7: Athra+Athra: P2 cuts to the center to get some useful movement in during a 7-frame period where he can't go above a certain height without the top-right enemy firing. 1-8: Athra+Athra: I tried a bunch of routes here, this one turned out the fastest. I'm not thrilled with it. 1-9: Athra+Verbal: The benefits of Verbal's faster projectile speeds and range come to light. 1-10: Athra/Gascon+Mars: Mass suicide. P2 actually kills his turret faster than P1 due to P1 having to dodge projectiles. Mars was originally picked so he would die on this stage, but due to a faster kill in 2-5, he stays alive. This doesn't actually cost frames. The boss has some WEIRD immunity behavior, sometimes you can hit it for 2 consecutive frames, usually you can't, and it looks like it's per-player. All I can say is every shot is taken at the exact frame it is for a reason. 2-1: Verbal/Verbal: Manipulated the last archer to a slightly better position. 2-2: Verbal/Verbal: The opening volley causes lag but the contact kills save time. P1 logjams enemies on P2's side to reduce wait. Manipulated an enemy on P2's side to avoid it moving up into his path. 2-3: Verbal/Verbal: The stage is bottlenecked on the top ambushes and it's faster for P1 to ignore the roof enemies so that they trigger as soon as possible. 2-4: Verbal/Verbal. 2-5: Verbal/Mars: The stage is bottlenecked on a single ambush chain and P1 gets first input. This leaves plenty of time for P2 to lose a character without costing any frames. I edited Mars' death in 1-10 out and moved it to here. 2-6: Verbal/Verbal: It's faster to shoot projectiles to the opposing side due to the weird angles.
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
GoodGen scans the (World) one as Gain Ground (J) [!], which is the only other Gain Ground ROM that GoodGen seems to recognize. It syncs with my WIP, so I'm not sure it matters. I guess I'll keep using (UE) since it looks like that's the one that was released in the US.
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
The game is Gain Ground. I'm having a hard time telling what the [c] actually means, some sites say it's "cracked" but GoodGen's docs say it's a "known faulty checksum," as opposed to [x] for a bad checksum that should be good I guess. If that means the cart shipped with a bad checksum, then I could see it being compatible with [!], and might mean finding it without [c] would be impossible. There isn't a published run, which is why I'm trying to start one. :)
Experienced Forum User, Published Author, Player (131)
Joined: 5/21/2012
Posts: 74
Location: Cary, NC
I have a similar question: If I have a dump marked (UE) [c][!], is that an OK dump to use? Gens auto-detects it as NTSC, would it be OK to run it as NTSC?