Personally, I’m thinking that if an infinite is detected the character performing it should automatically cancel into a special which causes knockdown.
I think that could work…
I can’t wait for this game, it looks like tons of fun.
Having the game automatically end your combo for you kind of goes against good game design, IMO. Forcibly ending the player’s combo by having them do a move they didn’t actually do is a step AWAY from the kind of freedom that fighting games should strive for. If the opponent doesn’t burst out of the infinite, that’s their fault. It was their fault for getting caught by the combo in the first place, so forcing a player’s combo to end is just giving negative feedback to them for doing something right.
I’ve played the game myself, and the infinite prevention system in place is great.
some info about the game
Wow that assist thing sounds very cool. I’m still on the fence about infinite detection, and I don’t like the character designs, but there is a lot of potential here
Heh, didn’t realize someone responded to my post. I suggested that because the player who gets stuck in the infinite actually gets rewarded under the current system. If, for example, the opponent techs right as the infinite detection kicks in, they’re almost guaranteed to be in a better position than if the player had simply performed a knockdown right before infinite detection kicked in. I’d say that a guaranteed knockdown is a pretty good reward, myself.
Is Skullgirls going to be a commercial indie game?
Edit: Nevermind, reading on home page. It’ll be cool to see this project released, really awesome work.
hopefully there won’t be any pc versions just console versions.
that way it will get most play and profit.
just don’t enable lan and use Steam
kthx
:tdown:
Honestly, this game looks gorgeous. The animation and character design is reminiscent of BlazBlue, Nightmare Before Christmas, and a 1990’s surrealist anime, FLCL. I think this is a potent design direction allowing for very dramatic and stylish Super Moves.
The infinite detection is strange but I think it’s a step in the right direction. I’m not a fan of the depleting hit stun found in Guilty Gear, BlazBlue, and now MvC3. I’ve heard it can make landing a Super Move at the end of a long combo very difficult. I like the idea of clear cut detection. If you can combo into and out of an infinite in this game it could make for very cool mind games as far as how many reps to do, and when to guard crush.
I just read about the custom assists and that is amazing. I would have never thought about that. It’s pretty brilliant.
Anyway, this game looks and sounds awesome. I’m almost more interested in this than MvC3 now. Hopefully we’ll see more characters and mechanics soon.
Also, I feel like there aren’t as many people excited about this as there should be. Is the word not getting out there like it should???
So I guess it’s a battle to see who can find the longest finite combo then. Finites, for the win?
bring this game out now!!
Yo is there a demo of this game available? I would love to just play the beta/alpha version against CPU or something.
This would be a day 1 digital download for me, I’d rep this game in my city too.
i cant resist just torrenting something liek this if it appears on pc.
i wouldnt play it properly on pc so i wouldnt want to buy it on pc.
i want some kind of xbox live version then i will buy it without doubt.
wow…
this game looks simply amazing…
I am really ampted to see people take initiative and make there own games, (especially 2d fighters keep 2-d alive:rock:)
This is definitely a first day cop for me. Im so inspired by this project. Much props and good luck to the SG team! Man I better learn how to program you guys are beating me to the punch.
Skullgirls the new darkstalkers??
MikeZ asked me to look at the build at Evo said he actually would want to hear my long ass rants about what would make a good training mode. If you’re reading this, sorry for the long delay in responding, but here it is, followed by a few random notes.
Training mode: TL;DR version: Xcopy VF4:Evo.
Step one: Frame by frame record and playback, with inputs recorded as their in-game representation (e.g. a given frame would be saved as ‘up/towards + weak punch’, not ‘up arrow, left arrow, Z’). The same code could and should be used easily for match recording and playback for versus games. Having the training dummy repeat the recording loop should be considered the very bare minimum for any training mode. For the more technical of us, having that recording be an easy to modify file, either with an in game edit or by using text files, would be a pretty big boon. For an ST example, if we could edit the recording perfectly, it’d be a cinch to have a recording of a perfect Gief tick SPD for practicing reversals.
Step two: Using multiple recordings. It would be very handy to be able to specify a number of recordings (5 would be usefull and more than enough for anything I could think of) and give a percentage weighting to each one. After a recording is over and about to loop to the next, use a random number and the percentage weighting to determine which of the recordings to run, and execute it. An easy example of this would be for practicing defense. In a marvel example, dummy set as Magneto, first recording would be dash, c.short xx c.forward, jump up, iad down roundhouse, sweep. Second would be dash, c.short xx c.forward, walk up throw. Third would be dash, c.short xx c.forward, cross up trijump, sweep. Fourth would be dash, c.short xx c.forward, walk up c.short, sweep. Four options for the person training to try and defend on reaction after the dash in lows.
Step three: ‘First opportunity’ settings. This would have to be done different from the record and playbacks described above. This would be a setting for the dummy character where the first opportunity to do a move, the character does it, and remains normal until the next frame they have a first opportunity. Crap, that sentence sucks. I’m mostly talking about having the dummy do reversals, preferably with a delay option. Set the ST Ryu dummy to ‘First Opportunity -> 0 frame -> Dragon Punch (jab version)’ and they’ll always reversal dragon punch after hitstun/blockstun/knockdown is done. Reset him in the air, and he’ll do it the frame he lands. Set the ST Ryu dummy to ‘First Opportunity -> 1 frame -> Dragon Punch (jab)’, and there will be one neutral frame after those same scenarios before he does the dp. Practice with Gief against that, and you’ll know for certain with your tick SPD is frame perfect.
Having all of the specials/supers as options should be a given, but having the normals, including the three jump versions, and possible iad versions (where the jump and iad are done fastest possible) would be icing.
Step four Add the data for multiple records like described in #2 to a single file, that also specifies the stage, the player and CPU locations (so they’d start in the corner if needed), the human controlled character, the dummy character, a specified amount of successes needed to complete (either ‘5 out of 10 attempts’ or ‘5 in a row’ type) and a set of success trigger conditions that is checked during each recording such “human character hit/not hit/taken damage/not taken damage/put in blockstun”, "cpu character hit/not hit/infinite detected/taken X hit combo/guard broken’, and probably others that you could think of better than I. Lastly, some explanation text so the player knows how to proceed and what to take from it. With that information together you have your character specific challenges ready to go. With the four way Magneto example above, the success triggers would have to be ‘human didn’t take damage && human put in blockstun’, with explanation text like "Blocking Magneto’s various attacks on reaction. Magneto will dash in with low attacks, and then either throw you, hit low, hit high, or cross up. Block the dash in lows, and then either block or beat out his followup attacks."
For the exercises practicing combos, set the recordings to empty, and have the success flags checked and reset whenever the dummy is in a neutral frame. For the exercises where the dummy is using a First Opportunity setting, the setting in the file and menu would have to be different from the recordings exercises, but the periods to check and reset the success flags would again happen whenever the dummy is in a neutral frame.
With that framework in place, it should be pretty easy to make a catalog of training exercises, and if the file format is easily editable, the community can help provide their own. If a two player recording could be put in as well as an example of how it should be done successfully, along with the ability to view inputs, more the better.
Random crap that I have no clue will apply at all to your game:
- Have the AI interface with the game via inputs, not by commands. It’s pretty annoying when playing an AI that has the ability to full screen walk up -> Sonic Boom. (If the communication between the AI and the main game code is done using inputs, and letting the game engine separately determine the result, then it would also make it possible for you to setup the AIs to combat each other with a genetic algorithm. If there are input bugs like CvS2 roll cancelling, I guarantee a few thousand or more generations of genetic Ai’s would find it, and you’d be able to see it in the match replays)
- Neutral hitboxes (where the character is vulnerable) should always be at least once pixel outside of the bounding box. Maybe I’m just salty about how many characters a Rufus dive kick will whiff on from point blank range; please ignore if using pixel by pixel hit detection.
- Please keep hitstun and blockstun values the same for the vast majority of moves that don’t knockdown. This is one of my biggest gripes about SF4. Having the blockstun be less than hitstun takes away from the aggression of the game in my opinion. If anything adding a few moves with greater blockstun than hitstun would be a good thing to throw in; the ‘universal overhead’ type moves in Garou:MOTW did this and I love that aspect.
- Online play should include a training mode options, both for playing the other player in training mode, and also to practice in training mode while waiting for a match. Sadly, SvC for Xbox1 is the only game I know of to do this, and waiting in a lobby with nothing to do sucks.
- Custom soundtracks of course. For PC versions, the ability to select an MP3 to play for each specific stage.
- Random Select
- Blind Select in versus.(pressing and holding a button hides the cursor, which can still be moved by the controller, releasing the button selects the character. Casual players will press and release the button after moving to their character and never be affect by the blind option)
I know I can be hard to read at times, so let me know if I need to clarify any of my ranting. Best of luck on the project, I’m eager to see where it leads.
You loved the fact that being hit by a Universal overhead gave the opponent an opportunity to super you? It was probably one of the poorest design choices I’ve ever heard.
I’ve played games of garou where I’d purposely not block an overhead so I could super my opponent. That seriously shouldn’t work in any fighter ever.
- Yep, precisely the point. A slow, telegraphed, move that allows the pressure to be kept if blocked, that can be kara canceled into air specials (or occasionally ground specials) depending on character, and usually puts you at a small disadvantage if it hits. The two factors that make for a good fighting game in my opinion are offense and speed, and you cannot get more offensive than taking a hit to gain an advantage.
- Your opinion, but not one I share. I grew up taking jump in hits so I could throw them in SF2; every single one of them had more blockstun for aerial attacks than hitstun. Consider your objection to my suggestion properly lodged.