This is false information. You have a set time (given in frames) to enter the next direction. The only (virtually) unpredictable aspect is that the game uses frame skipping to increase speed, so a few frames will not be shown and inputs will naturally not be considered.
I thought that it was pretty well-known that moves are not guaranteed unless they’re inputted within a certain window? In other words, it works somewhat on a probabilistic scale from 100 to 0%? Is this not the case? If it is, then my question is why would someone include that in a fighting game?
That’s not the case. Let’s take a Tiger as an example. As soon as you’ve entered ↓, you have 6 frames to enter 
. After that, more 6 frames to enter → and 10 to get a Jab version, 9 frames to get a Strong version, and 7 frames to get a Fierce version. If you do not complete the motion, the timer resets and you have to try again.
So you’re saying that the input window is consistent for all moves in ST? Even Sirlin says that this is not the case. In fact, he even mentions on his blog that one of the things they “fixed” about Akuma is his “random input window” for his teleport. If he says it was fixed, then they must have known it was not consistent in the first place. I think that pretty much everyone knows by now that in some cases in ST (regardless of frame skip) some moves would just not come out because of internal counters or whatever. It’s also easily testable. The technical details about why it happens are irrelevant.
My question is about why this type of input system was added into the game. What were they trying to achieve, or is it more likely an unintended side-effect.
You can’t be serious.
It’s to add more risk to the nearly catch all of downback.
Probably because Capcom’s coding was so wacky back in the day, even if the general design ideas were very good. (Nowadays is usually the complete opposite, at least as long as you don’t clash a fireball with Rolento’s knife…)
Hahaha, that is so wrong, lmao.
It adds more strategy value, and it also adds a risk to both sides (crouching/catching a crouching character). Since you need to know how to capitalize a crouching character, and the defender need to be aware of not getting caught or he would eat a combo that can potentially hurt a lot.
It works as Pasky explains. There is a timer, it gets decremented every frame, it has the same value when you start the motion, then it gets the same when you continue it. If you take too much time to continue the motion, the move might not come out, since the timer will end. This will only happen if you keep using the same direction for a long time, such as walking DP, or Hadou from a long time in neutral crouch, instead of down-back, which is what everyone uses.
I got no idea what Sirlin did, what I know it how the game works. But Akuma: I believe no-one bothered to check his inputs. It is easy to test, and Born2SPD tested all those commands (but Akuma’s) and uploaded them to the SRK Wiki. Sirlin also through Cammy lost to Dhalsim and nerfed one of her moves. It turns out she actually loses that match in ST.
Lots of things have been discovered after HDR was released. Several things people did not know have now been unveiled, such as how Claw loses his claw.
It was not added into the game. RAM memory does not lie.
I liked SF2 when I had a 1-in-200-something chance for a special move to come out when I pressed a button.
And if you hold a direction for more than the alloted window for that input and you input the next direction at the same moment the counter is reset, the first direction does not count toward the input for a motion command, and therefore the move will not come out. To the player, this seems random. This is all documented in that thread that you linked me to (I read it many times before and tested some of these things myself).
I’m not sure what the point about arguing the technical details are. This topic is about design, not implementation. If it was unintended that the input windows be inconsistent, then so be it.
No, the coding is still just as bad.
OK.
That’s a question for which we know the answer: it just seemed a way to get it done, when they were coding SF2. And it was a way that worked miles better than SF1 - and significantly better than Fatal Fury 1, 2 and Special, BTW.
I think the problem is that you keep using terms like “inconsistent” and “random” when in actuality (at least, as far as I can tell from what oldschool_BR is posting), neither of these is the case. It’s just more demanding than it should have been.
Honestly though, I think discussion of anything from before roughly 1995 (or probably later, honestly) should basically be off the table for purposes of this thread - because the answers for “Why would they do something like that?” in games from that era are generally one of “They weren’t really sure of the repercussions”, “That was the best way they could come up with at the time” or “They were pretty much just trying stuff, because no one had done anything like this before.”
While I’m sure there are interesting answers for some of the questions, stuff specifically relating to games that are like eighteen years old puts us firmly in the territory of “even the people who were there and made those decisions probably don’t remember exactly what problem they were trying to solve.” territory.
Creativity and not quality. I have a high level of respect for the stuff they did, not that it ended up the most polished. I don’t like Garou that much but they went pretty wild with different ideas. That lead to a lot of better mechanics.
Hijack all you want 
It’s an open conversation, and I’m not some utmost authority. If you have a good answer, share it.
The things I specifically stated are added risks of crouching… yes. You figured out the puzzle
For hitstun values, there’s a few things to consider: Higher hitstun adds more emphasis on a hit. Stronger hits have higher hitstun so they feel like they do more damage. The use of time is critical to pointing out that hey, this is more powerful, and makes for a pretty observable feel difference. Higher hitstun allows for easier hitconfirms as well. Pushback is similar, think of a battering ram vs a punch to the gut. More power = more movement. That said, I think pushback has a really high amount of emphasis on combat flow. If you want an example of a game with TOO MUCH hitstun, look at Blazblue CT. A lot of hits take forever, and the game feels sluggish because of it.
Competitive players that know a lot about game design are… really useful. To say the least. There are really good players that are clueless on the matter and would contribute very little other than finding observable problems. The good players will break your games, the good players who know things about game design will tell you why they’re breaking it. They’re amazing for QA -and- design and a lot of companies have people on gameplay teams that are really, really good at the game they play. But yeah, you can always count a competitive player in as a good tester more often than not. They will break your games harder than anyone else, but they might not tell you what feels good.
I think nerfed hitboxes are actually due to time constraints. Almost all the boxes in SF4 and Marvel look like they were made for getting hitboxes done faster. you’ll see they trace the outlines of the characters in SF4, and in Marvel use cicles at the character’s joints, which means you put in a few numbers and bam, you’ve got a blob that resembles the character. The size might even be generated by the distance from the camera, haven’t checked. Arc System hitboxes are pretty much the same as they’ve always been. Not sure about other games! 3S actually had hitboxes as “nerfed” as SF4, btw.
There’s one distinct advantage to the smaller hitbox sizes: it looks more consistent. With the boxes conforming to the character limbs, when two hits interact, it doesn’t look like they hit some random space out of nowhere, which looks wonky and feels “bad” to players who just wanna punch dudes.
I think people took offense to your “adds arbitrary difficulty” comment.
It does! I don’t think the gain is worth the complexity increase, but it’s a pretty marginal thing.
The probabilities of missing certain moves due to whatever reasons have been calculated and documented. For all intents and purposes, they can be considered random and/or inconsistent. It’s much easier than saying “I missed the input because I happen to press X on the 6th frame of the 3rd input motion”.
Like there isn’t a lua script to help you get the motions down for ST now.
Spoiler
[media=youtube]rjGmWWXV3f0[/media]
The problem comes that many people like to call “arbitrary” barrier at anything that they just happen to not like.
Is not arbitrary when there is a well thought reason behind of why we have it on the game, and more importantly, when it adds to the strategy of a game.
It was his misuse of “arbitrary”.
Shouldn’t be surprised by it though looking back through the topic.