This thread has a link to a Reaction speed test that mimics defending overhead attacks in a Fighting game. Some of you may have seen this test before, its fun. The test is good because you have to click ONLY when an overhead attack comes at you, and not when a low or medium attack comes.
There are 2 overhead attacks that might be performed, one is a cartwheel kick, the other is a big yellow thing. You have to defend 5 overhead attacks which come randomly interspersed with normal attacks which you must ignore, and then it calculates your average. If you click when it is not an overhead attack, then you lose all your progress for that set of 5, and you fail the set. There is an attack where the sprite stabs with her hair. Its annoying because it is bright yellow and the animation extends quite high on the screen, so its easy to think its an overhead attack by mistake. So be careful about that one.
Please, no cheating. Please post all of your stats, and not just your best one. This is really important.
So before you start get a notepad up.
You can have one trial go to get the hang of things. But after that, please post ALL 5 of your scores and the average for each set that you do.
And record how many sets you fail. Its quite addictive, and you can certainly improve your score dramatically with practice.
So for example: 1st set (trial so ignore) 2nd set 35.4 34 36 35 Average 35
3rd set 28 27 28 29 28 Average 28
4th set 26.3 28 Failed Failed 26
5th set 22 18 20 24 21 Average 21
Please write down how many times you failed. And then ofcourse share your best times too!
Its really important that you share your worst result (your 2nd try or atleast one of your earliest tries) aswell as your best. And how many times you tried overall too.
Once you are finished, there is one last thing I would like you to do. Redo the test but this time dont do it as fast as you can, but instead do it at a pace that you find 'relaxing but still snappy'. Try to find the pace that you personally find to be 'fun while still keeping you on your toes'. For example, try to imagine you are playing a hypothetical new RTS game, and you are microing your units. You see the enemy units raise their weapons, they are about to shoot and you must react by pressing dodge left. Try to work out what you would find to be a fair, comfortable and fun reaction speed. NOT your fastest possible.
Bit of info for people who like this sort of thing.
1000 milliseconds in a second If a game is running at 60 frames a second, then each frame lasts for 1000/60 = 16.666666 milliseconds. If a game is running at 120 frames a second, (and your screen is 120Hz) then it would be 8.333333 milliseconds.
The test in the OP is 60fps. A modern ‘esport’ that is being developed as of 2015 should be 120fps in my opinion though.
An average persons reaction time is roughly 250ms plus or minus 20 I would say. This is pure reaction, doesn’t include anticipation or training which would occur over time in a computer game. Note it also doesn’t include choosing a response from multiple options. It is simply detecting a change.
Many people are slower, their reaction time is around 290 milliseconds. Some people are slower still. Sub 350 would include most people. Sub 400 would really safely include most people. Sub 500 includes everyone. Fastest reaction time is 120-140 milliseconds.
Its important to note that the above test is not even close to a fair test because we are all using vastly different hardware; inputs of mouse/keyboard, the computer itself, and the output monitor. With that caveat, Im gunna say: A good average is around 21 frames. Up to 25 average would be pretty good too, especially on first try. To get below 30 frames average is quite easy. To get below 35 is certainly easy. To get 18 consistently might be possible. To get lower might be possible too but extremely hard.
So for game design, we have to choose what is a comfortable speed and what is a more skilful speed: If a game was balanced around 35 frames for reaction time, this is a comfortable number, still fairly crisp but not stressful. 21 frames is a bit stressful and might feel a bit luck based. That said, 23 frames is actually not that hard to get, especially with just a little practice. The developer of a game has to plan for the skill improvement of the players over time.
The ideal mechanic would be that experts can execute attacks that give the opponent less time to react to. But noobs can only execute attacks that have a long time to defensively react to. Eg1, a noob picks up the game for the first time and presses attack button. It’s the most simple to execute attack in the game and therefore it is also the easiest to defend against. The attack has a long animation period in which it can be blocked. Eg2, an expert is playing, and he does a complex combo attack. The number of inputs is higher, and requires precision timing to chain together. The output is an attack with a smaller animation period in which it can be blocked.
This means that experts can play the game at a faster pace, but noobs can also enjoy the game. Low skill floor, high skill ceiling. Its similar to how in tennis, a noob cant hit the ball hard or accurately to attack. Similarly this weak attack makes it easier for his noob opponent to defend by getting the ball back. Tennis experts on the other hand are able to hit the ball fast and accurately, so the attack is much stronger. But at the same time an expert defender is faster on his feet than a noob and has faster reactions so he can return the ball.
If there are 60 frames per second then: 30 frames = half a second = 500 milliseconds 24 frames = 40 percent of a second = 400 milliseconds 18 frames = 30 percent of a second = 300 milliseconds
Its important to remember how the actual animation itself is incredibly important too. Change of silhouette, change of colour, change of position are 3 of the main things that the eye can detect. It takes atleast 50ms for the human eye to see a change, and atleast 50ms for the brain to process and send impulse to the finger muscles. But this is not taking into account the actual animation itself. Consider a character that does 2 very similar punches and uses exactly the same animation for the first 8 frames, and then on the 9th frame it clearly divides into a 'red hand punch' or a 'blue palm slap'. Lets pretend the defender has to press crouch to dodge the red attack but jump to dodge the blue attack. So even though the attack started 9 frames ago, the person cant react until much later, even if his eye has detected an incoming attack from frame 1.
----------EDIT ADDED RESULTS---------- The following is posted on page 3: Here are the stats. Some people posted 1 result, some people posted three results, and one person posted 6 results. Of the people that posted more than 1 result, the difference over time was not significant. So I have simply calculated the mean average of all the results. One person posted his ‘relaxed’ results. It was the same speed as his other 3 results at full concentration.
Raw data (each line is a different person) 25.2, 21.9, 22 30.4, 26.7, 27.5 23, 21 23.3, 27.7 25.2, 25.5, 27.4, 26.7, 26.5, 26.4 26, 31, 27 24, 22, 22 24.8 23.6 28.6 28, 28, 27 21.9
Sum = 740.3
29 total bits of data
740.3/29 = 25.53 frames
Mean average = 25.53 frames
Now lets work out what that is in milliseconds.
We previously worked out that if there are 60 frames in a second then 30 frames = half a second = 500 milliseconds 24 frames = 40 percent of a second = 400 milliseconds 18 frames = 30 percent of a second = 300 milliseconds
To work out the 24 frames example: 60 frames/24 frames=2.5 100/2.5=40 which is 40 percent of a second. So that’s 400 milliseconds.
Putting our newly calculated average of 25.53 frames into the formula: 60 frames/25.53 frames=2.35 100/2.35=42.55 which is 42.5 percent of a second. So that’s 425 milliseconds
Lets compare this number to the other reaction time test where you only have to react to a change and do NOT have to also choose a response. The average reaction time in the other thread here: http://www.teamliquid.net/forum/general/481373-whats-your-reaction-time was around 250 milliseconds. This would represent getting 15 frames in the OP test.
So no surprises; reaction time is significantly slower when you also have to choose a response. This is a much more applicable statistic for video game design. 425 milliseconds, not 250 milliseconds, cool.
Thanks for the help guys.
This little experiment has helped back up my thinking that for a future hypothetical realtime strategy esport, you should design around a flow between actions of half a second/500 milliseconds/30 frames (in a 60 fps game) or indeed 60 frames in a 120fps game. (Ofcourse you can still allow faster players/professionals to play much quicker)
Press a button to start. Wait for an attack. If not an overhead attack, don't press anything and continue waiting. If an overhead attack (flip kick or hairball), press any button. Repeat x 5.
If the game is stuck for some reason, I suggest just pressing buttons until the 5 trials are over.
I would also suggest playing it a bit to get used to the animation of the flip kick and hairball, as well as the other attacks, so that you are on a more even ground. Fighting game players would have a bit of advantage in picking up the animation cues and have an advantage in the reaction time.
ok you dont click . why did you say you click? "all you have to do is click" . i was sitting there like a spastic waiting for something to happen then got like 4 faults
fault 24.2 24.7 22.6 29.2
18.6 24.3 22.8 fault (double press) fault
fault 19.4 748 24.5 fault
yeah i cbf with this
played Soul Blade (the original Soul Calibur) so much when i was a kid i was able to go infinite in survival mode (turned it off after 108 kills)
this simulates that basically, which is fucking cool.
so yeah if you like this go download Soul Blade and a PS1 emulator
the game is like broodwar quality, literally top 3 games of all time
the story mode has like 3 unique endings for each character, and during cutscenes your character is presented holding your selected weapon, something i dont think you see in games 10 fucking years later. oh yeah and there are like 10 unlockable weapons per character
Haha, I was clicking as well after pressing spacebar... first trial was like 40 seconds.
Also, don't use spacebar. Use an arrow key or something, the spacebar is heavy and adds to the delay.
My thoughts are: Flip kick is much harder for me to react to. Jumping hairball attack is much easier to react to, but I get it confused with the jump back low split kick thing.
i find it disorientating coz i want to move around and just block everything.. i mean its cool but im like .. i need to fucking play soul blade like a pro now and not ffuck around with this 1 button shit
edit: dude since you're so interested in game mechanics you should definitely check out Soul Blade (playstation 1) coz its a great example of this done to perfection imo
This feels completely irrelevant to fighting games. If you're up against a player, or even an AI, there's clear signs that an attack is coming. They'll be comboing, or you'll be playing footsies and dancing in and out of range.
There's an extreme difference between that and waiting 10 seconds to push a button.
players do their best to mask their intentions so you are left with 1 thing to react to , the attack animation.
soul blade is amazing because its longest combos are 4-5 hits and the animations are designed in such a way that a good player is able to recognise and block all of them.
it is not a "mechanically challenging" game (you don't have to press fast or time your presses) but instead comes down to mental trades with your opponent where you are trying to get each other to fuck up your decision-making whilst both being on a timer (you can only block so many attacks before you are unable to block for the rest of that game)
On April 07 2015 07:11 FFGenerations wrote: i completely disagree
players do their best to mask their intentions so you are left with 1 thing to react to , the attack animation.
Of course they do.
But you're not sitting in front of them for 10 seconds and waiting for them to push an attack button. Reacting to specific moves in a game of footsies is a vastly different experience than watching two motionless characters in a corner with randomized inputs with massive intervals in between.
theres no difference.... if you are able to discern attack animations , regardless of whether you are "playing footsie" or standing still next to your opponent, then you will be able to block them
it is a different experience to a fighting game of course but it is still the fundamental mechanic that they are built from. if the player is unable to discern attack animations within an appropriate timeframe then the game becomes one of chance
in a well designed fighting game you CAN just stand there like a lemon and block every attack that comes your way (if you are good)
On April 07 2015 07:24 FFGenerations wrote: theres no difference.... if you are able to discern attack animations , regardless of whether you are "playing footsie" or standing still next to your opponent, then you will be able to block them
it is a different experience to a fighting game of course but it is still the fundamental mechanic that they are built from. if the player is unable to discern attack animations within an appropriate timeframe then the game becomes one of chance
Well, maybe it's just me, but my problem with these kind of "reaction test" games is never "oh no, it's so fast, I can't react". It's "fuck this is boring, I'm going to start thinking about other things".
On April 07 2015 07:24 FFGenerations wrote: theres no difference.... if you are able to discern attack animations , regardless of whether you are "playing footsie" or standing still next to your opponent, then you will be able to block them
it is a different experience to a fighting game of course but it is still the fundamental mechanic that they are built from. if the player is unable to discern attack animations within an appropriate timeframe then the game becomes one of chance
Well, maybe it's just me, but my problem with these kind of "reaction test" games is never "oh no, it's so fast, I can't react". It's "fuck this is boring, I'm going to start thinking about other things".
yeah well i said that in my first post already lol :D
Right, my 4rd try (I really didnt get it on the first 3, kept failing.) 3rd try: 30.4
5th try: 26.7 6th try: 27.5
It seems I should be able to get sub 30 consistently. All the different attacks are super distracting though especially when a single test can take either 30 seconds or TWO MINUTES because of the massive delay between overhead attacks sometimes.
Haven't really seen improvement yet but I'm fairly happy with my score!
On April 07 2015 06:52 FFGenerations wrote: played Soul Blade (the original Soul Calibur) so much when i was a kid i was able to go infinite in survival mode (turned it off after 108 kills)
this simulates that basically, which is fucking cool.
so yeah if you like this go download Soul Blade and a PS1 emulator
the game is like broodwar quality, literally top 3 games of all time
the story mode has like 3 unique endings for each character, and during cutscenes your character is presented holding your selected weapon, something i dont think you see in games 10 fucking years later. oh yeah and there are like 10 unlockable weapons per character
i seem to react slightly slower to the flip kick than to the hairball
seems the flip kick has a longer animation time and my brain feels like my response is "timing" to the end of the longer animation when it hits the character
1st try average 25.2 2nd try average 25.5 3rd try average 27.4 4th try average 26.7 5th try average 26.5 6th try average 26.4
My single lowest was 18.4 on the hairball and my single highest was 37.4 on the flip kick. I think I failed like 7 times. I think the animations are very misleading and it actually put me in a bad mood.
I only did one practice run, as suggested. Here are my next 4 tries (did not fail): 24,23,41,25,28 , avg 26 27,34,34,32,29, avg 31 27,24,23,34,26, avg 27 Relaxed: 27,29,30,33,29, avg 29
Basically, anything 25 and below is the hairball, 26 and up is the weird kick flip. Relaxed = me playing shadow footsies until I get the signal. The thing I don't like about this test is that, usually, you switch to block high when your opponent stands up/hops, but in this case, it will cause you to fail. I'd really like a test where it only matters if you're holding down on the first hit frame.
Maluk wrote: Average : couldn't even block one hit lol.
I think you must be doing it wrong because it doesnt matter if the Overhead attack animation finishes or not, you still record a number even if its like 1000 or more.
@Pwere, thanks for the stats and for doing it relaxed too. Im not sure what you mean by 'playing shadow footsies until i get the signal' because you cant move the character around. But either way, if you mean Relaxed = doing the test in a chilled out way and not as fast as you possibly can, then ideal. Thats part of what I am looking for. Your relaxed numbers are essentially the same speed as your full concentration numbers which is kind of funny.
Thanks for posting your stats everyone, keep em coming.
Maluk wrote: Average : couldn't even block one hit lol.
I think you must be doing it wrong because it doesnt matter if the Overhead attack animation finishes or not, you still record a number even if its like 1000 or more.
@Pwere, thanks for the stats and for doing it relaxed too. Im not sure what you mean by 'playing shadow footsies until i get the signal' because you cant move the character around. But either way, if you mean Relaxed = doing the test in a chilled out way and not as fast as you possibly can, then ideal. Thats part of what I am looking for. Your relaxed numbers are essentially the same speed as your full concentration numbers which is kind of funny.
Thanks for posting your stats everyone, keep em coming.
I meant that I was thinking about what my character would be doing while keeping an eye out. Relaxed enough to still have fun and not feel stressed, relaxed muscles, etc.
However, that number is misleading because I got mostly hairballs, and I was getting better at that point. I'd say I was about 4f slower on average, which is still surprisingly low for how much better it felt.
I meant that I was thinking about what my character would be doing while keeping an eye out. Relaxed enough to still have fun and not feel stressed, relaxed muscles, etc.
Yes thats ideal, thanks.
Also, you say you were getting better at that point. This is certainly true, and its actually a good thing because it simulates a real online game eg starcraft where you get more accustomed to scenarios through repetition and anticipation etc.
I have collected all the stats you guys have shared and will do a little summary in a few days. This is your last chance to share your results, consider this the final bump.
Hi, D3v from Shoryuken.com here. Was pleasantly surprised that something made mostly for the FGC is being talked about here.
That said, this needed a few clarifications.
On April 01 2015 07:44 CardinalAllin wrote: Bit of info for people who like this sort of thing.
1000 milliseconds in a second If a game is running at 60 frames a second, then each frame lasts for 1000/60 = 16.666666 milliseconds. If a game is running at 120 frames a second, (and your screen is 120Hz) then it would be 8.333333 milliseconds.
The test in the OP is 60fps. A modern ‘esport’ that is being developed as of 2015 should be 120fps in my opinion though.
All fighting games (or at least the good ones) are locked to 60fps. These games are basically state machines and alot of data is locked to specific frames. Even in modern, 3D animated games, specific data is still tied to each individual "frame" such as hit boxes, hurt boxes, invulnerability, armor, etc. Also, attacks are still timed in terms of frames (start-up, active frames, recovery) and alot of the math to determine if something can combo (a second hit can be applied before the opponent comes out of hit stun) or is safe (recovered before the opponent comes out of block stun) is calculated based on frames.
Seeing as Teyah's test was made using a 2D fighting game as a base, there's a reason it only does run at 60fps.
On April 01 2015 07:44 CardinalAllin wrote: Its important to note that the above test is not even close to a fair test because we are all using vastly different hardware; inputs of mouse/keyboard, the computer itself, and the output monitor.
Since we're working with "frames (60ths or a second), you can find a baseline based on this. Basically, a true "tournament spec" setup is one where the input gets processed in less than 1 frame, or 16.67ms (plus the natural 1-3 frames of lag inherent in processing, which the program should already account for).
The FGC has already put up alot of resources for this such as DisplayLag.com.
That said, unless you have a really, really bad setup, (high input lag monitor plus laggy controls) you shouldn't be getting anything over a frame of lag on your computer.
Hello D3v. What a pleasant surprise. Post your stats!
I see now that the way I juxtaposed these 2 sentences makes it look like a possible criticism of the test:
The test in the OP is 60fps. A modern ‘esport’ that is being developed as of 2015 should be 120fps in my opinion though.
That wasnt the intention though. The OP is not about Fighting games in particular. Its about game design. I know that present day Fighting games are locked at 60.
not even close to a fair test because we are all using vastly different hardware
I knew when I wrote this it was gunna come back to bite me. I was too dramatic.
You reckon that the hardware differences will result in only 1 frame of lag. Thats interesting. Can you tell us about the travel time of the keys as your finger presses down. It is my understanding that this alone is enough to add several frames. A question, does the internet lag affect results much?
The whole reason I made this thread is because the test you link is a different kind of reaction. It only tests detecting a change and not choosing the correct response.
The test in the OP is 60fps. A modern ‘esport’ that is being developed as of 2015 should be 120fps in my opinion though.
That wasnt the intention though. The OP is not about Fighting games in particular. Its about game design. I know that present day Fighting games are locked at 60.
60 is probably as good a baseline as any since not everybody is going to have the hardware to run at 120fps, what with most games just hitting 30.
More importantly, a game that you can bump to 120fps (in other words, not locked to 60fps), isn't based on frames like fighting games are (so the code runs regardless of frames).
not even close to a fair test because we are all using vastly different hardware
I knew when I wrote this it was gunna come back to bite me. I was too dramatic.
You reckon that the hardware differences will result in only 1 frame of lag. Thats interesting.
Any system will have a native amount of delay that you can't really do anything about. For example, the old CPS2 arcade board that ran Street Fighter II had around 4 frames of native delay.
More importantly, most modern games are can be designed so that the amount of native delay doesn't really change. For example, Street Fighter IV on PC doesn't really change the amount of delay no matter what the hardware is, as long as it's within the minimum recommended specs (the sole exception is that V-sync adds 1 frame of delay, but even that can be controlled).
So the only things that need to be taken into consideration are the delay on the controllers and on the display.
The fastest monitors have about 09ms of lag. Meanwhile, at least based on our own testing, most controllers can go as low as 03ms (with an averayge of 06ms) of delay so that's within the 16.67 (not 12.5 that I mistakenly put down earlier) for each frame.
On April 27 2015 19:05 CardinalAllin wrote: Can you tell us about the travel time of the keys as your finger presses down. It is my understanding that this alone is enough to add several frames.
On April 27 2015 19:05 CardinalAllin wrote: Of course there's a slight amount of delay, but this is why we use gear such as Sanwa buttons or Cherry switches with very, very low actuation force, to minimize the amount of delay.
The numbers I quoted above, for 3-6 ms for controllers, are tested using manual input though.
A question, does the internet lag affect results much?
Yes it does, which is why fighting games are still best played offline. Though the bigger issue tends to be variable delay, which is input delay that changes as the ping time between two players changes. This is why stuff like rollback netcode has been developed as we believe that it's more important to keep the delay fixed (and let small rollbacks handle desyncs) than screwing up people's timing by slowing the game. Too bad not many devs are using it.
oh i see i suppose to press a key asap after the spinning cartwheel and it doesn't even matter if I block it before it hits me? I stood around after got hit by it thinking I'll just fail this one but no, I press something and got like 800 frames LOL
why the hell would you call that an overhead... okay trying again.
you should disable mouse click as a click... so I can actually copy/paste the text here instead of writing it down, I tried to copy and everything disspeared.
okay there's another overhead attack, i see... most of mine are faults tbh haha... the cartwheel is pretty hard to queue in, I started looking for the pantie shot as a queue, but maybe there's a better one
evanthebouncy! Thanks for posting your stats. I will include them in my results.
D3v: I agree most people dont have the hardware for 120fps. The main thing being a 120Hz monitor. They are just coming down into the good value for money window, and probably in 2 years time they will be affordable for people looking to upgrade. Other hardware considerations are ofcourse the actual grunt of the machine itself to process the frames. If a developer wanted to start making a new game today that was similar to Street Fighter IV then I agree that 120 would be unwise as a design goal due to the average gamers specs and monitor. If I were to be very very general, I would say most games run at 30 on consoles but most games on PC can be run at 60 on the average 'enthusiast' spec machine (not necessarily with all max settings). And yes, current games that can run at 120 frames are mostly Shooters that are not lockstep.
You keep mixing professional setting vs us doing an online reaction time test using random hardware and using the internet. May I quickly just repost the whole paragraph for context:
Its important to note that the above test is not even close to a fair test because we are all using vastly different hardware; inputs of mouse/keyboard, the computer itself, and the output monitor. With that caveat, Im gunna say: A good average is around 21 frames. Up to 25 average would be pretty good too, especially on first try. To get below 30 frames average is quite easy. To get below 35 is certainly easy. To get 18 consistently might be possible. To get lower might be possible too but extremely hard.
The only reason I put in the caveat was to cover my arse from looking silly if I had just said 'A good average is 21 frames' all on its own. But anyway, from your second post it is now clear that we both agree that for doing this test here at home on our various hardware and over the internet that it is far from a fair test, and that there is a plus or minus window of atleast 4 frames.
Regarding my very loaded sentence:
A modern ‘esport’ that is being developed as of 2015 should be 120fps in my opinion though.
This is for a 2d lockstep game. The Esport game I personally envisage is a brood war RTS type of thing, designed specifically for the professional setting that you have described; tightly controlled high spec hardware, LAN connection. If I were to begin developing a 2d fighting game today, I would personally also make it 2d and 120 frames per second. This isnt to say that I deem 60fps games as less skillful or anything like that. Random fact, Im using a model M keyboard from 1991 and a rollerball mouse without a scroll wheel (seriously someone send me a spare mouse).
Netcode; the choices made for how to implement this is vital ofcourse. If I were to start making an Esport game today I would make it for LAN first and foremost as I previously mentioned. However, it would be wise to develop an online version too otherwise the game would probably not ever get popular. It is important to consider it a separate version entirely in my opinion. I agree with you that the best solution is to "keep the delay fixed" rather " than screwing up people's timing by slowing the game."
Overall, the discussion in the thread comments has concentrated on hardware. Its a critical element to consider when designing a game, especially an Esport which is what this thread is mostly concerned with ultimately. There were several other topics brought up in the OP such as: -Comfortable speed (for beginners and relaxed sessions) VS Top skill speed (for best gameplay at professional level) And how you can design to accomodate both and all inbetween on a scale. -Animation considerations (again for best gameplay at professional level, not just eye candy)
@D3v thanks for sharing the juicy details here such as:
The fastest monitors have about 09ms of lag. Meanwhile, at least based on our own testing, most controllers can go as low as 03ms (with an averayge of 06ms) of delay
Its great to read the views of a Fighting game expert such as yourself so cheers. Game designers have much to learn specifically from the Fighting game scene (in many areas not just hardware). I myself have a considerable amount of notes from Fighting games Ive studied.
Its important to note that the above test is not even close to a fair test because we are all using vastly different hardware; inputs of mouse/keyboard, the computer itself, and the output monitor. With that caveat, Im gunna say: A good average is around 21 frames. Up to 25 average would be pretty good too, especially on first try. To get below 30 frames average is quite easy. To get below 35 is certainly easy. To get 18 consistently might be possible. To get lower might be possible too but extremely hard.
The only reason I put in the caveat was to cover my arse from looking silly if I had just said 'A good average is 21 frames' all on its own. But anyway, from your second post it is now clear that we both agree that for doing this test here at home on our various hardware and over the internet that it is far from a fair test, and that there is a plus or minus window of atleast 4 frames.
What we really need to find out is how many frames of delay Flash has, plus whether or not things like browser choice affect it. Then we can get the baseline level of delay for the system. That said, with as thorough I I know Teyah to be, that's probably already considered in his calculations.
On April 28 2015 06:37 CardinalAllin wrote: Regarding my very loaded sentence:
A modern ‘esport’ that is being developed as of 2015 should be 120fps in my opinion though.
This is for a 2d lockstep game. The Esport game I personally envisage is a brood war RTS type of thing, designed specifically for the professional setting that you have described; tightly controlled high spec hardware, LAN connection. If I were to begin developing a 2d fighting game today, I would personally also make it 2d and 120 frames per second. This isnt to say that I deem 60fps games as less skillful or anything like that. Random fact, Im using a model M keyboard from 1991 and a rollerball mouse without a scroll wheel (seriously someone send me a spare mouse).
120fps for a fighting game might cause some problems though, especially with how certain data is calculated by the community. The bigger issue though why no one is trying to do this (even the more forward thinking indies) is that since most things are tied to frames, doubling the amount of frames means double the work and more chances for making mistakes.
That said, it would be interesting if someone made a fighting game that ran at 120fps, but that simply repeated the data from one frame to the next (so what would take one frame, would now take 2 frames).
On April 28 2015 06:37 CardinalAllin wrote: Netcode; the choices made for how to implement this is vital ofcourse. If I were to start making an Esport game today I would make it for LAN first and foremost as I previously mentioned. However, it would be wise to develop an online version too otherwise the game would probably not ever get popular. It is important to consider it a separate version entirely in my opinion. I agree with you that the best solution is to "keep the delay fixed" rather " than screwing up people's timing by slowing the game."
Sadly, some devs still don't seem to agree.
On April 28 2015 06:37 CardinalAllin wrote:ssional level, not just eye candy)
@D3v thanks for sharing the juicy details here such as:
The fastest monitors have about 09ms of lag. Meanwhile, at least based on our own testing, most controllers can go as low as 03ms (with an averayge of 06ms) of delay
Its great to read the views of a Fighting game expert such as yourself so cheers. Game designers have much to learn specifically from the Fighting game scene (in many areas not just hardware). I myself have a considerable amount of notes from Fighting games Ive studied.
Our genre, due to having been codified during the 90s, is one that's most sensitive to things like lag and such. The move to HDTVs has actually caused alot of issues once it was discovered that modern stuff is slower than old school CRTs. This is why once we learned this, we started looking more and more into sources of input delay.
After about 10 practice trials because I couldn't figure out the controls and what I was supposed to be doing.
24 22.2 19.3 22.2 21.9 Avg 21.9
I probably could get better if I could distinguish between Millia's hopping low kick and the overheads better. The low kick leads to a lot of false positives for me.
Here are the stats. Some people posted 1 result, some people posted three results, and one person posted 6 results. Of the people that posted more than 1 result, the difference over time was not significant. So I have simply calculated the mean average of all the results. One person posted his ‘relaxed’ results. It was the same speed as his other 3 results at full concentration.
Raw data (each line is a different person) 25.2, 21.9, 22 30.4, 26.7, 27.5 23, 21 23.3, 27.7 25.2, 25.5, 27.4, 26.7, 26.5, 26.4 26, 31, 27 24, 22, 22 24.8 23.6 28.6 28, 28, 27 21.9
Sum = 740.3
29 total bits of data
740.3/29 = 25.53 frames
Mean average = 25.53 frames
Now lets work out what that is in milliseconds.
We previously worked out that if there are 60 frames in a second then 30 frames = half a second = 500 milliseconds 24 frames = 40 percent of a second = 400 milliseconds 18 frames = 30 percent of a second = 300 milliseconds
To work out the 24 frames example: 60 frames/24 frames=2.5 100/2.5=40 which is 40 percent of a second. So that’s 400 milliseconds.
Putting our newly calculated average of 25.53 frames into the formula: 60 frames/25.53 frames=2.35 100/2.35=42.55 which is 42.5 percent of a second. So that’s 425 milliseconds
Lets compare this number to the other reaction time test where you only have to react to a change and do NOT have to also choose a response. The average reaction time in the other thread here: http://www.teamliquid.net/forum/general/481373-whats-your-reaction-time was around 250 milliseconds. This would represent getting 15 frames in the OP test.
So no surprises; reaction time is significantly slower when you also have to choose a response. This is a much more applicable statistic for video game design. 425 milliseconds, not 250 milliseconds, cool.
Thanks for the help guys.
This little experiment has helped back up my thinking that for a future hypothetical realtime strategy esport, you should design around a flow between actions of half a second/500 milliseconds/30 frames (in a 60 fps game) or indeed 60 frames in a 120fps game. (Ofcourse you can still allow faster players/professionals to play much quicker)