I never said your math or calculations was wrong. It’s amazing that you could ever get a degree in anything with such poor reading comprehension.
Your data is a farce. You are lying about the nature of your data. Stats on tainted or skewed data doesn’t mean shit.
You know what fine, I dare you to walk up to any stats professor and ask him if I am doing anything wrong. I think you will not like the answer.
Using your data I separated the offline and online into two different categories and performed the proper tests to see if you were right (which the data suggests you weren’t)
These are the same tests that you would use to determine based off of density of a metal was pure or not. You would prepare the metal using a known purification method and perform several density measurements. You would then perform the same measurements on the metal in question and perform these tests. If the density was proven to be statistically different then you would know that the metal of interest was not pure. Likewise if the two are determined to be of the same population then the metal is pure. Aka there is no measurable difference between the metals from two different processes. They use this in quality control.
Hell this is the method they used to discover argon. They found two different methods were proven statistically to be giving two different results and as a result they knew the gas they recovered from the air was different from the gas that was chemically generated.
What you fail to understand is that this is the method you would use to see if your statement was right or in this case can we say for sure that the Cammy Akuma match up offline is completely different from the online match up. You are right in saying there are differences between Casual and Tournaments and that there are differences between offline and online. What you don’t understand is this does not directly imply that the match up is any different in these scenarios. It could be 4-6 no matter what setting we are in.
Again at this point you just keep proving you don’t know anything about the math and are just giving unproven opinions.
So you are telling me that any stats prof will agree that you can take data from apples and take data from oranges and if they are similar declare them as the same?
Really?
So you are telling me that statistically there is no difference between winning 40% of the time and 60% of the time. Because that was the data I was looking at.
All that needs to be done is count the wins and losses and compare them to see if there is a difference and how significant that difference is.
And here I thought i was proving your complete and utter dishonesty.
No they would tell you I am checking to see if we are comparing apples to apples or apples to oranges. The flaw in your logic is you assume the match up is different just because of a 2 frame lag. The 2 frame lag effects both players equally so in the long run it may net no change. At the same time it could very well cause a change in the match up and give one character an advantage.
I am telling you that with the data size there is an error associated with it. As I gave you an example before, flipping a coin 7 times and getting heads all 7 times is perfectly possible. It is not however the correct mean of the population and as such for only 7 flips the error associated with it is rather large.
That is why the t-test and F-test were invented in the first place because not everyone can take sample sizes of thousands of measurements so they needed a reliable way to prove if two groups were the same or different. Again go ask any professor, they will tell you the same thing.
If you got a sample size of 500 matches then sure you can do that because the error for each sample size will be small enough that it is easy to eyeball it and tell. On the other hand given our current sample size there is still a sizable error because of the fact that in real matches crap happens; however, by averaging more and more matches you get better and better results because random factors are…well…random. As a result the random factors eventually cancel each other out in the long run.
I’ll give you another example. Let’s say after 6 coin flips we have a perfect 50/50 distribution. That is 3 heads and 3 tails. If we flip the coin once more we are guarenteed not to have a perfect 50/50. Instead we will have a 57/43 distribution. This is 7% higher than the true average of 50%. But according to our t-test, this result does not indicate our expectation of 50/50 is wrong. It instead says that while 50/50 is the expected value that the real value will deviate from the expectation value. That is why we use what is called variance to measure the expected error of our prediction. We expected to have about 3.5 heads, but the deviation says we realistically can received values between 1.8 and 5.2 heads 60% of the time and as it turns out we have 4 heads and 3 tails which is within this interval. Therefore our population of 7 flips is actually within reason. Actually for 7 flips you realistically can have 0-100% heads because the error is so large for only 7 flips. But to get 100% on 500+ flips is realistically hard as hell. You would end up getting something closer to 50/50
You know you are still stuck on the idea that just because online and offline have differences that the match up can’t be the same for both. Again go ask a professor.
Let me use your logic somewhere else. I have a new experimental drug that cures cancer and I want to test it out. I hand out the drugs to one group of people with cancer and at the same time I hand out sugar pills to another group of people with cancer. Now according to your logic I can not compare the two because one group has a miracle drug and the other has a sugar pill.
Well the question is what happens if the drugs don’t do what we designed them to do? If that is the case then there is no difference between the two groups because neither group received any benefits and as expected the two groups should have identical distributions.
The same can be applied for what we are talking about. Let’s say for sure there is a 2 frame lag. “Does the lag give one character advantage over the other?” is the question that pertains to the match up. Does it screw both characters over equally? Do the players even notice it?
I mean for that matter we also need to separate PS3 and Xbox because it has been proven there is a timing different between the two systems. Do Cammy players get an advantage by playing on a monitor with a slight lag?
No they wouldn’t.
We already know we are dealing with apples and oranges. What you are doing is finding “similar” stats between the two and using that as a reason to declare they are all apples. That is wrong.
In this particular match Akuma needs to be reactive to Cammy pressure and needs to be tight on her during wake up pressure. Adding more frame delay makes both of those tasks more difficult. Harder to DP jumps and dive kicks and harder to safely attack her wake up.
That’s why I said more data needed to be collected. The data pool is too small.
I understand, but we aren’t trying to determine random factors or unknown quantities, this is simply a collection of wins and losses. The more you collect the better.
That’s because it’s a fact. They are two different games that are played differently. The inherent delay and factors of lag cause players to alter how they approach the matches. Different combos, different tactics, etc. These changes changes the matchups.
You can compare anything, however you cannot say that each drug is the same drug.
Just because something is similar statistically doesn’t mean it’s the same thing.
These are the questions that require the data to be collected separately. That is where you will find the true answers.
If you find there is a large variance in results of online vs offline then you can confirm that something is up and vice versa.
Players generally prepare for tournies by practicing on the system of use. While there is a difference between the two I haven’t seen much evidence that pegs it at more than a frame, maybe less.
But yes, all categories could be separated and calculated and compared to determine the differences if any.
Not sure but I don’t know of any useful tourney that uses laggy equipment.
Anyway, now that we have this discussion back to a civil tone, if you want to continue we should either have our own thread or through PM’s. We are bumping too much relevant stuff lately.
Loyal I really appreciate your effort, your patience is admirable. The thing is, you’re just wasting your time. Shin is right.
Yea I don’t even know why I am bothering.
No instead it means there is a 95% or a 99% chance they are the same thing. Pretty good odds if you ask me. Seriously get a damn probability book, take a class, or do something besides run your mouth on a topic you aren’t qualified to talk about.
facepalm
facepalm
shuts down computer
Ok, now that we all agree that Shin is right, let’s talk about something more related to the thread.
So I’m constantly getting beaten by Dan, his backdash and super taunt are so godlike and op. What can I do in this matchup? How can I deal with the constant taunts? Akuma’s toolset isn’t enough
hahahahahahahahahahahaha!!!
On the whole LoyalSol vs ShinAkuma204 thing, I won’t bring up statistics because any kind of mathematics is my weak point. But, I will say though, to Shin, (I’m not looking for an argument by the way)
Do you really, truly and honestly feel that Offline vs Online (Japanese) play is like comparing apples to oranges?
I mean, it just seems, if that were the case, why do Japanese players play online constantly? Why are they constantly in Endless Battle playing? Why did Tokido and Daigo even play 100 straight ranked matches in the first place? I know it wasn’t to get GrandMaster status online for shits and giggles.
If it was as drastic as apples to oranges, then why would they risk playing online (apples) when they have to go the arcade the next day (oranges) ? It seems that all their online training would mean jack shit when they hit up the arcades. (When we know it really does mean something)
(For the record, and I promise I’m not trying to drop names here. lol, but I have a lot of top Japanese/Asian players on my friends list, and they are ALWAYS in endless, ACE E I RI N, robin01, Playerjun52, Bullcat, etc, and I’m assuming it’s to level up their game (which includes offline play) … And on the whole Daigo said that Japanese Net Play is almost identical to offline because “Well we have Daigo one time saying that it was virtually identical to offline when promoting an online tourney his company was sponsoring.” Yeah, but that doesn’t mean it’s not true, you seemed to totally discredit him saying that because of a sort of “he HAD to say that”, which is kinda silly, it’s not that far out, Air has told me a few times actually that he couldn’t tell the difference of online play to offline play, when he was in Japan, and he also said that many top players felt this way, including Tokido himself. Lol)
Sorry for the mega paragraph. 
U fail to realise when i told u to stop bringing out my name for your random arguing , this wasnt the first time i visited the akuma forum . Btw I dont need some random theory fighter or random dude from these boards to tell me what set is an indicator of a matchup, like you know what you talking about or he was there or something . Someone asked me about my opinion , i answered that’s it.Your arrogance still amazes me.No wonder top players dont post on srk forums anymore.
This is a good discussion :tup:
LoyalSol, how does your statistical analysis account for the fact that it is possible to change your approach to the match up depending on the environment (increase in lag due to online, monitor, game console, whatever…)?
Isn’t there a difference between the following:
Statement 1: Akuma vs. Cammy results in the same probability of winning offline and online.
Statement 2: Offline is equivalent to online.
It seems like the statistical analysis can only be used to support Statement 1.
The ability to change your play style depending on the environment prevents the statistical data from supporting Statement 2.
If I was someone who only played offline I would want the online matches to be removed from the calculations. I would not care about the approach to the match up that results in success in an online match.
I am legitimately asking this question, I am not trying to argue with you. I think that this is interesting.
Well online is always different. Sure it can be close, but it will still be different. And hey maybe I’m wrong, and if that is so collecting enough data would show that.
I was never under the impression it was your first visit, so I’m not sure what you’re talking about.
I’m not trying to convince you of anything. You want to take casuals as the indicator of the matchup despite the tourney results, that’s your perogative. I just offered my take on it.
This comment is pure trash. You come in here, accuse me of something I never did. Get exposed for it and you don’t even bother to apologize for it? And you want to talk arrogance? LOL
I wonder what this thread is gonna talk about when AE drops and Super Cammy dies 
It’s probably still gonna be mad entertaining.
I don’t know why there’s such a hard-on for matchup numbers in some character forums. This argument has at least been going on since like last September or something - I don’t really give a damn if I’m name dropped at all or whatever (dem e-points), but surely Akuma has other matchups?
Why don’t you guys just agree to disagree, lol.
Also, just to let you know, Cammy-anyone is 12-0. Put down the stick. She she she so good.
He does, he just dominates them so hard we don’t need to talk about it.
That’s because you can’t DP her divekicks. HURRY UP AND FIX THAT AE!
We’ll start bitching about yun/yang once AE drops and it’s probably gonna be even worse, hahah. Other than the 2 twins akuma will probably dominate all the other matchups and that’s good
To this day I am still saying “WHAT THE HELL CAPCOM?!”. When I think of the twins. They almost feel like Seth without the low health sometimes because they have everything.
Yeah man I know. I don’t know where the video is but this guy (I think on this forum) posted a video of Yang’s super which didn’t even hit until half the timer had gone and still managed to take out 60% health, I’m shuddering thinking what it will do to Akuma, just the thought of it makes me cringe!
The twins make so little sense.
Capcom - “We’ll nerf Cammy and Rufus and Seth and Ryu, but we are gonna replace them with EVEN MORE BROKEN CHARACTERS!”
WTF?