While reading today’s installment of AIGameDev [Side Note: If you are interested in AI development for games, this site should be on your "required reading" list], which, while not generally relevant to the main topics of this blog, I did find some information that might be of interest to many members of the community. In particular, today I came across a section called "The Reality of Human Reaction Times", and was immediately enthralled. Consider the introduction to the section :
Human reactions times vary depending on what type of reaction we are talking about. The reaction time to a simple sensory stimulus ranges from 150-300 ms (i.e. 7–15 frames). Note that this is simply for the reaction to a stimulus. ANY reaction. For example, an enemy wouldn’t react to a muzzle flash from a gun (“That’s a gun shot!”) for about a fifth to a quarter of a second. If you add in a simple deliberation factor (“Is he shooting at me?”), reaction times are significantly longer.
I’ve had any number of discussions with other members of the community about human reaction times, network ping times, and latency with regard to Combat: Samurai Island, and how those affect combat. Of those factors, the only one intrinsic to a person, and the only one improvable by practice alone, is reaction time.
I’ve long suspected that the best C:SI warriors have an ungodly reaction time, and the aforementioned article pointed out an easy way to prove or disprove that idea :
If you want to really get a grasp on how slow our reaction times are, here’s a graph based on this online twitch test. Seriously, try taking it yourself!
Naturally, I had to head off straightaway to take the test myself, and I’m quite pleased with my results. I’m currently suffering an injury that makes it phenomenally difficult to move my right arm, and even twitching my fingers on that side hurts more than I’d like, but I averaged a decent enough score (203ms) to put me in the top 50 people in the last 24 hours (all-time scores put me to shame, though).
I’m wondering, if we could somehow convince all the top players to take this test, would we see a strong correlation? I’m betting that we would, and would in fact be tremendously surprised if we did not, though that might be very instructive.
In addition to that, if you read the user comments and other information on the site, it highlights the importance of training to improve reaction times. This could easily be extrapolated to demonstrate the viability of some of the available third-party C:SI training tools.
So how about it? Are you game? Take the test and post your results!!!
Popularity: 62% [?]
I think there is some sort of diminishing returns principle at work with reaction times. Take for example trying to slash someone when they kick you. As long as you are quick enough to get the slash in before they can block again, being quicker won’t really give you any additional advantage.
The one thing that has improved my fighting more than anything else has been learning to predict my opponent’s next move. If I have to wait until I realize that they are kicking and not slashing, I have found that it is almost impossible to get the slash in before they can block. But if I am confident enough that they are about to kick to risk the penalties if I am wrong, then I can react immediately when I see them lower their block. In fact, you can often anticipate exactly when they will lower their block, so you don’t even have to wait for that. I strongly believe that reaction time alone, no matter how fast, can only make a person an average to above-average fighter.
Anticipation of your opponent is powerful, no doubt about it, but part of me still clings to the notion that the ability to quickly process data, anticipate what’s coming, and react appropriately within a very small window of opportunity is still tied to having good reaction times in much the same way that real-world martial artists must study the physical basics enough to firmly implant them into muscle memory before they can spontaneously and effortlessly act in accordance with the situation : It’s foundational, not the epitome of skill.
I agree that reaction time is a required foundation to have. I just think that there is very little advantage in improving it past a certain point.
That may well be true, though I’m still very curious indeed to see whether there is a clear correlation between fast reaction times and C:SI performance. Speculation is all well and good, and very entertaining, but I’d *LOVE* to see actual data
For what it’s worth, my average was 194.8 (currently number 41 on the top 100 as of this time), with my quickest time being 173ms. I haven’t dueled in quite a long time, but I used to be fairly handy with a sword. Not sure if that’s useful to you, Robby, but hopefully it can be the start of a wonderful new data-mining project for you! =D
I don’t know if it’s useful either, but it *is* interesting. Especially as I consider how you used to handily lop off my head when we fought each other
P.S.: Yes, I know there are other factors at play with me vs. Naeko, but I’ll thank you very much not to mention how poor I am at combat strategy :p
I had an intersting discussion about reaction times a few years with a cognitive researcher. We were discussing reaction times in baseball, and apparently, when you factor in the speed of a fast pitch, there is no way that the hitter can hit the ball (apart from luck) by processing the information on trajectory AFTER the ball leaves the pitcher’s hand and THEN starting a swing. Human reaction times are just too slow. The implication is that great hitters anticipate where the ball is going to be from reading the body language of the pitcher as he winds up. That is a tacit skill acquired through experience, and some people just naturally are more able to develop this skill. I suspect CSI is the same. We anticipate by reading the patterns of our rival, and by making the appropraite decision. The reaction time of good fighters is fast, primarily because they are anticipating before the event what their opponent’s next move will be. Indeed, to some extent they dictate their opponent’s next move by their own last move. Its not simply about reaction time, and its not simply about abticipation, the two are joined at the hip. They are inseperable.
I took the test and my reaction time was about average. It was last night and I don’t remember the exact number. I think it was around 115, 120. I agree with Judo 100%.
Ok, I finally took the test and got 206.9 average.
I wonder whether I could come up with a way to measure reaction times and anticipation in C:SI? Before anyone starts objecting, I mean on a strictly voluntary, opt-in basis. I’m just curious, is all, and I’d love to explore such things further. I just am not sure of the proper way to go about it.
Um…my reaction time was about 260-270 average, on the low end of those reported thus far. But that’s enough usually to counter effectively. I guess I’m not terrible with a katana either. I struggled with countering and tanking for a long long time before I learned that there are only a finite number of ways to deal with any C:SI action. So I limit my focus to countering only those possible attacks. I think that was my way of learning how to read patterns and to break a spar down into mini-encounters. I guess my results would tend to support Colin’s assertion that reaction time, although important due to the limited timeframe of an attack animation, is not the end-all-be-all factor in determining C:SI success.
Yeah, I agree that it’s not the most important factor, but I did think it was one of the most testable ones, and very interesting besides.
I think the test would be fascinating, because I think there is a good chance that the data would reject the hypothesis that reaction time is correlated with success. Incidentally I took the test. My reaction time was horrible. Average of 318. I suck.
Well Judo, the last time I had the opportunity to converse conversationally with a cognitive psychologist was about 15 years ago
If you have any thoughts on how to conduct such tests in SL, I’d be fascinated to hear them!
Judo’s reaction time might have been terrible, but her Spar-Fu is strong.
I am surrounded by psychologists! Most of them are nuts (I am not of course ;-)). To test, I guess one of the training tools could be adapted to measure average ms reaction time, place the tool at popular combat sites (Meiji, Edo, etc) and ask people to try (is that a random sample? probably not), record the data, and correlate with their CSI record (which I know is an imperfect indicator of skill, but the best we have). Best experimental design I can come up with at 9:15pm after two glasses of wine!! Now back to watching American Idol. Oh the horror.
I think you’d have to log both users’ key strokes with a timestamp that is ideally at the millisecond level. That doesn’t sound like something that would be very reliable in the environment of SL. If you are really serious, maybe write an external program to log the events that volunteers would run on their computers when they sparred. But then there’s still the problem that the two programs running on each persons’ computer would have to have their clocks synchronized to within a millisecond.
Tracking two consenting user’s keystrokes at millisecond resolutions is possible in Second Life. I have some code that I use quite frequently that performs exactly that function (for one person).
Correlating them might be more challenging, though.
*wanders off in thought*
Oh yeah it was 215, 220, not 115, 120 for me. xD
249.8 for me
/me cries
250.6 … I’m slow? :/
vint falken’s last blog post..SL musicians Get It Together
Slower than me
WhooHooo!!!! I’m faster than *TWO* people now!
In all seriousness, I’d obviously hoped to see a stronger correlation, and I still believe that I would in RL, but there doesn’t seem to be a strong connection for SL.
I strongly remember one summer years ago, during an Aikido seminar where the founder of our school (Minoru Kurita sensei) came over from Japan to teach us, how profoundly impressed I was by his reaction times. When we would just begin our attack with the bokken, he would already be moving, already know what we were going to do as we began the strike.
He’d shouted “First rule of Aikido: GET OUT OF THE WAY!”, then proceeded to show us how, if your mind were calm and still, you could react spontaneously and appropriately no matter what your opponent does. If you are not where the opponent’s sword will strike, you do not have to block, and can then use his committed action against him.
The result looked magical to me, so quick were the reactions, and so perfect.
Often times I’ve seen people watch a video or live demonstration of Aikido and make comments that “they’re faking it” or “that ukemi is doing all of the work”, and while there may be more or less truth to that during public demonstrations and among various schools, if they’d ever had the truly awesome opportunity of first-hand experience of attacking a master martial artist like Kurita-sensei they would surely show much more respect.
.
One thing I found with the test was how much I’m not used to trying to time a mouse click quickly. I don’t do games and C:SI was a huge struggle at first, so at least for me I think any speedy reaction I do have is with my left hand on the attack keys. With C:SI the mouse click is coordinated with what you want to do, but doesn’t necessarily have to be all that well timed, at least compared with the attacking keys.
Oh, and I agree with Judo about the prediction end of it, and how you know if you show someone one thing, they are most likely to follow with some other thing. I think a lot of people are faster than me at reacting to things from moment to moment but if I can possibly still get more wins its by being a tad more clever with how I set and change pace, switch up patterns, and predicting theirs.
Digging up an old thread since my comments aren’t getting Spamified anymore. LOL So here’s a dive into the Hyperkitty brain. xD
So first raw numbers. Urm…averaged around 200. So I’m not slow. But I’m not fast. But…
I *definitely* agree with Kasumi on the not being used to a twitch test. Like the only way to explain it is that there’s no flow…no rhythm…it’s like my first spar in the evening when I get home from work and I get destroyed. xD
Cuz like my brain sees the twitch test flash and just goes WTF like I don’t know what to do. It’s like deer in headlights before I figger out “oh yah, move my finger”. But in a spar I get into a flow and rhythm with both hands and react faster than my twitch test.
Now Judo on the baseball analogy thing totally explains my style. Like scary explains it. Cuz thinking about it made me realize exactly how much my style relies on reading my opponent as opposed to reaction time. O.o
Watch me fight. Like watch me close. Really close. I’ll have what appears to be *insanely* quick reactions when countering. Watching my YouTube videos…especially Brittany #2…I land slashes in like a 100ms window where she unblocks and goes back into block. Like my death blow to her in that video is a slash where she unblocks for a *split second*.
No way I coulda seen that. No way I coulda reacted to that. Period.
Watching the rest of the Brittany, Straight, and Shindo videos like shows something else. My opponent will be literally right in front of me and I’ll turn away or look up into the air and slash at nothing (I ATTACK THE DARKNESS LOLZ). I’m slashing at air.
Why? Cuz that’s where I *thought* they were gonna be. *sneeze*
Aimee Congrejo’s latest blog post is : Leaving Yakuza and going clanless for a while