Rate Their Chances Returns! Day 194: "Predicting the Direct"

[jaytalks]

I think you need to step back for a moment and look at what you're doing. You're writing a dissertation on a mathematical model that essentially just boils down to "sales are good/bad, but these other factors matter too so I guess this rating is valid." It's like that whole mathematical cult of people who spend their lives predicting and tracing patterns in every thing imaginable even if it's really just based on the whims of people.

Seriously, we can make model of correlation all we want, but at the end of the day everyone here is already doing that without some needlessly complicated mathematical system. Our brains are already designed to detect patterns through observations, and your model doesn't seem to be doing anything different. And even then the fact remains your model of sort is based around figuring out the mind and decisions of a clearly random individual. You yourself mentioned before how Wii Fit Trainer "surprises" your model. So how in the name of Arceus is that reliable?
All of this doesn't make you sound more objective or informed, at least to me, it just comes out as all-out obsessive and absurd.

And I don't say this out of spite, I seriously just happen to be concerned about you. You're crossing the line into guy-from-the-movie-Pi territory. Like I said before, I normally saw your posts as the beacon of reason in the middle of absurd percentage ratings, but now everytime I read what you say I can't help but think of WTF IS GOING ON.

I think you are looking at it from the wrong perspective. Sure, those of who don't use data are making the same type of predictions, but not with the data behind it. It's comparable to pundits who were making predictions about that 2012 election. But there were individuals like Nate Silver who used data that ended up being right. It's no different than when someone cites a quote and analyzes it for the purpose of their analysis. I'm not saying Erimir will be right similar to Nate Silver when it comes to predictions, but that there is merit to analysis and using statistical data.

I'm not speaking on behalf of Erimir, but in the name of my (admittedly limited) statistical knowledge. Even if something has a 14% chance of happening, if that said thing happens, that doesn't disprove the model. It had a 14% chance of happening. Furthermore, if you read further on in the analysis:
"The model is surprised though, because Wii Fit is, well, not a conventional choice and the trainer is not a playable character. If not for the genre and such, it would've predicted more like 1.2 slots. That said, from what I can tell, this is higher than most people would've rated Wii Fit, so I don't feel too bad about the model being wrong if it was less wrong than most."

It's surprised because WFT is not a conventional choice of character, it's genre, and the lack of playability of the character. He even says that if you removed those factor, it would have had a slot. So he explains how the "surprise" happened, a little bit later than in the post. The model is stronger than how most would have predicted WFT prior to shkler reveal at E3. Much stronger. So if anything, WFT actually shows the strength of the model.

 

[Erimir]

Seriously, we can make model of correlation all we want, but at the end of the day everyone here is already doing that without some needlessly complicated mathematical system.

Because I'm just doing it for the hell of it?

Also because I think it's a good check on the tendency to overrate.

And even then the fact remains your model of sort is based around figuring out the mind and decisions of a clearly random individual.

People keep saying that.

I don't think Sakurai is random. He might want to throw us some curveballs every once in a while, that doesn't mean he's random.

You yourself mentioned before how Wii Fit Trainer "surprises" your model. So how in the name of Arceus is that reliable?

Because that's just one instance? It's not guaranteed to be right in every instance, especially when Sakurai does do something that doesn't fit the overall pattern. But that doesn't mean there's not a pattern.

It's also less surprised than most people were. It gave about a 15% chance of Wii Fit getting a representative. How many people would've rated WFT that high?

All of this doesn't make you sound more objective or informed, at least to me, it just comes out as all-out obsessive and absurd.

And I don't say this out of spite, I seriously just happen to be concerned about you. You're crossing the line into guy-from-the-movie-Pi territory.

Ummmm... not really. This isn't even really that hard. I had already collected the sales data, the rest was just loading it into the software and trying out different combinations of features.

I'm just a dude in a computer science program who took a machine learning class and is wasting time. Just because to you fooling around with statistics doesn't seem like something someone would do without being insane doesn't mean it is so. Clearly you're not into math stuff as much as I am. I took multivariable calculus, AP statistics, discrete math, number theory and math modeling in high school. I mean, I've done things for fun like come up with my own spelling system for English, creating an "alien" alphabet, compiled data about gay rights to make graphs (showing the amount of population living in countries with legal gay marriage), etc. I took a machine learning class last year, so I'm just applying it to Smash Bros to see how well it can do. You should see what some of the other people in my school do for "fun".*

It's just me being really nerdy. Do you think I'm attaching some super importance to this?

Like I said before, I normally saw your posts as the beacon of reason in the middle of absurd percentage ratings, but now everytime I read what you say I can't help but think of WTF IS GOING ON.

Well ok... I was explaining the math part to make it understandable since you said you didn't get it.

*Actually, there was a guy who used machine learning techniques to teach a computer to play Super Mario Bros

http://www.youtube.com/watch?v=xOCurBYI_gY

 

[XenothiumX]

Anyway...

Linear regression is not new at all. And an 81% correlation is pretty damn good. Even the 66% correlation you can get with sales figures alone is good. Claiming that sales don't matter in the face of that just makes you look mathematically illiterate.

But you're right, Sakurai is both totally random and unpredictable and it is foolhardy for me to think that an 81% correlation means anything. But also Sakurai is practically guaranteed to put in a character you like from a game he worked on, despite leaving Kirby with only one rep for two iterations even though Kirby has 10x as many sales as Kid Icarus. Cuz that makes sense.
Points for Mugatu references!

I would recommend making a graph based on your data for showing your reasoning than explaining with a bunch of words.

 

[FalKoopa]

Dr. Kawashima:
Likelihood: 0.1%
He is based on a real life person, which is a problem as you'll need his permission. Not just that, Dr. Kawashima appears on other consoles as well (though the floating head & hands is exclusive to the Brain Age games). I don't think he has a good chance when we already have Wii Fit Trainer and will likely get Miis as well to represent the casual side of Nintendo.

Want: 0%
Being based on a real life person is a huge negative for me.

Rosalina (honestly, we should have rated her before the likes of Omastar or Yarne or Tharja)
prediction: 15%
I think her reveal as a playable character in SM3DW and the Mario Galaxy stage will skew her ratings.

Dr. Mario x 5

 

[Glaciacott]

Stuff

Alright, understood. If it's that simple and entertaining for you I see why you'd go ahead and do it, especially about something like Smash which you enjoy. I guess to me so much mathematical modeling just seems like a sign of something being wrong, but I understand to you it's as valid a form of argument as writing essays would be to someone else.

But yea, like Xenothium says, graphs are your friends. Had to take too many statistic courses in college that I developed the habit of completely powering down upon seeing statistical models, and I very much would still like to understand what it is you're saying. For example, what I initially had understood was that the model gave WFT a 1.5%, and I'm also unclear as to how viability of the game to fit in a fighting game is factored into a percentage.

In any case, let's leave this for your other thread since it deals with that, and it's completely unrelated to this thread.

On topic: wonder how much the Mario Galaxy stage will factor into the Rosalina ratings. People seem to be getting more and more hyped for her on miiverse, although it being miiverse kind of implies that means nothing.

 

[Aqua Rock X]

Chancellor Cole outrandomed Omastar? Bwahahahahahahahahahaha!!!!!

Dr. Kawashima

Chance: 0.01% - Pretty much explained by everyone else.

Want: 0% - As a playable? Nah, I think he'd be better as an event mode specific character with Fox, Falco, and Wolf. He can bring his hands (and horns) along with him too.

Rosalina Prediction: 24% - As Dr. Kawashima falls flat on his face, Bluepikmin turns red! But amidst the shadows, a blonde female figure appears! Find out what this mysterious star-wielding lass is up to next time on episode 99 of Rate Their Chances: "Rosalina's Spotlight".

Nominations:
Bandanna Dee x5

 

[MasterOfKnees]

Dr. Kawashima:

Chance: 5% - I don't really see it happening. I'm not neglecting it, so I'm still giving him 5%, since he's not completely unfamiliar with Nintendo, and he does have sales behind his back. I don't expect another WTF character though, and if anyone, if won't be another one from more recent times I'm sure.

Want: 0% - I don't like him at all, he is creepy and would be way too weird for me. I do have limitations.

Rosalina Prediction: 18% - Many people are probably gonna rate her too high in the midst of the SM3DW hype.

Nominations:
5x Brawl Pokémon and Mewtwo

 

[BluePikmin11]

Love the conversations here, even though it's mostly negative now.

 

[XenothiumX]

Dr. Kawashima
Likelihood: 0.01%

Want: 1%
Only because I would like to beat the crap out of his face!

Rosalina
Prediction: 22.5%
Revealed to be a playable character in Super Mario 3D world and a Super Mario Galaxy stage?! I might need to raise the ante.

Nominations:
Krystal x5

 

[Smashoperatingbuddy123]

Dr kawashima

Both likely hood and want 0.01%

NOPE not going to happen

Rosalina

Prediction 45%

Was was going to say 5% at first but that was until the mario galaxy stage got revealed today that helps her to get in in playable characters

 

[BluePikmin11]

Prediction 45%

Was was going to say 5% at first but that was until the mario galaxy stage got revealed today that helps her to get in in playable characters

45% might be inflating too much for someone who just got to be a playable character, I would suggest lowering the percentage.

 

[colder_than_ice]

Dr. Kawashima
Chance: 0% - I just don't see how he could be playable.
Want: 0% - HE'S NOT WELCOME HERE!!!!

Rosalina prediction: 9.4%

Nominations: Aeron x5

 

[PK_Wonder]

Kawashima 0% due to legal issues and comparative lack of relevance to WFT, our current WTF-winner.

predict Rosalina 6%

Travis Touchdown x5

 

[BluePikmin11]

I'm really curious on how the ratings will boil down to.
From the many zeros here to the many 25%s and 15%s, I wonder.
Maybe 4.6%.

I totally can't wait for NIntendog's day.

 

[Seraphim.]

Dr. K

Chance: 5%
Want: 50%

Rosalina: 32% Recently was shown playable in an upcoming game and her potential home stage was shown in today's pic of the day, there will be hype.

Lucina x 5

 

[Cheezey Bites]

Dr Kawashima:

A true modern gaming star. The main character(? atleast the figureHEAD *badumtishh*) of one of the breakaway hits of the system and a game (kinda) that has had a huge impact socially. Just the fact that people talk about their brain age (ocassionally now-a-days, but still) or really understand the term, regardless of their lack of neuroscientific knowledge, is testament to how huge this game is, and Nintendo SPD spear-headed his popularity... but with his work being in books, on PC, iPhone and Xbox Kinect as well as DS make him less a Nintendo star and more a star of gaming in general... not to say that's stopping him, as he started with NSPD, he's best known on the DS, and we got snake (and PSASBR got Big Daddy, which has less to do with PS than the good dr has to do with Nintendo).

The big problem comes that he's based on a real person. You can butter it up that he's a fictional representation, but I feel Miis are opening up a can of legal worms, Kawashima is worse... and what happens if he changes his mind or doesn't renew his contract... we get mr dream all over again?

Sorry, blue, he deserves it, but I don't see it.

Chance: 0%
I think he's ineligable as being a real person (representation or no)

Want: 30%
I think he'd be a better AT, but the series deserves some serious representation in the 3DS version.


Rosalina: 12%
I think people are smart enough to realise this is too late to make a difference, and moreover that she's an unlockable 5th character, not a full character. We all knew she was popular (fro some reason) anyways, so it's no surprise they're wheeling he out for the fanboys... reminds me of Shadow for Sega.... Still, some people will rate her ludicrously high because relevancy is so important to smash... that's why it still uses classical music, because it's relevant and modern...


Tails*5
(Something something flying, believe, like him... I can't remember the lyrics...)

 

[Louie G.]

Did Blue even vote?
He's just replying to everyone's ratings.

 

[BluePikmin11]

Did Blue even vote?
He's just replying to everyone's ratings.

I did vote already?

 

[Louie G.]

I did vote already?

I don't think so...

Edit: Aw, wait, you did.
60% and 100%.
60%? Really? Okaaay...

 

[BluePikmin11]

Its a yes or no decision, so that's why its at 50/50 it can happen or not happen.
The additional 10% comes from the moveset potential most of you guys will never see.

 

[Louie G.]

Its a yes or no decision, so that's why its at 50/50 it can happen or not happen.
The additional 10% comes from the moveset potential most of you guys will never see.

Oh I see it.
It just isn't interesting.
Moveset Potential =/= Inclusion

 

[BluePikmin11]

Yes I know, it's only just the part of the inclusion of a character.

 

[McDuckletts]

Dr. Kawashima
Chance: 0.9%
Kawashima's saving grace is that the Brain Age is in fact a very popular Nintendo-owned franchise that currently spans over 3 games. But that's where the positives end for the good Doctor, unfortunately. Kawashima is indeed a real person, but most of you already knew that. Unless the real Kawashima is totally fine with them putting his character in, I can't see any real possibilities for him. Another problem is that I don't think Kawashima actually wants his character in Smash. His games promote having a healthy brain that can solve problems faster and have a better sense of memory. Putting Kawashima in an admittedly mindless fighting game would contradict what the Brain Age games are all about. While I don't think he's completely out of the question for a playable character, he's still extremely unlikely.
Want: 50%
A floating head that attacks people with math? Count me in.

Rosalina prediction: 25%
I feel Rosalina will do moderately well tomorrow. Maybe. I don't know. I'm just glad we're rating her before Pauline and Capt. Syrup.

Nominations
Banjo-Kazooie X5

 

[BluePikmin11]

Looks like Groose is not here to update yet. What's going on?

 

[Pacack]

You know what, change my chance for Dr. Kawashima to 7%. The maths won me back to the pessimistic side.

Make want 33%.

Rosalina prediction and noms are still the same.

 

[Deleted member]

Looks like Groose is not here to update yet. What's going on?

He's helping Link take down The Imprisoned with the Groosinator right now! You gotta give him some time!
In all seriousness, I have no idea.

 

[BluePikmin11]

Oh well, there goes the average.

 

[Golden Icarus]

Dr. Kawashima
Chance 1.5%
I don't really have anything to say that hasn't already been said.
Want 50%
I'm actually pretty interested. I'm just not sure if I'd want him over most other characters.

Rosalina 10.5%
We'll just have to see if her 15 seconds of fame is enough to boost her chances.

Nominations
Dark Samus x5

 

[Capybara Gaming]

Dr. Kawashima:

Chance: 10% - Casual, Popular franchise, but I don't expect him.
Want: 100% - floating head that fights by Teaching? Color me WANTING!

Rosalina: 10%. She's got a REALLY good chance now.

Nominations:
Takamaru x5

 

[Gam3rALO]

Dr. Kawashima
Chance: 5%

Want: 2%

Prediction for Rosalina: 1.57%

Nominations
Paper Mario x5 (I hope he can cut into a nearby day)

 

[FinalStarmen]

Edit: Aw, wait, you did.
60% and 100%.
60%? Really? Okaaay...

Did you actually expect any less?

 

[BluePikmin11]

Looks like the Groose is loose to end his day now!!

 

[Gunla]

He could have some connection issue or something. Either way, the day isn't done yet!

DAY NOT OVER​

GROOSE MIA​

 

[BluePikmin11]

Perhaps we could discuss how Sakurai can let Dr. Kawashima bypass the originated from game rule, because Brain Age is quite an important part of Nintendo's history, he would think Kawashima translates well as a character. Although any character not related to Nintendo at all would not get in, like Bugs Bunny for example. Never mind then, but it's something I would like to point out.

 

[Groose]

So, I had a big post typed out. But then, I had a crash and it disappeared. Woot!

Kawashima's Chance: 2.5%
Pros: Obvious uniqueness, comes from major franchise, obvious choice for franchise representation.
Cons: Difficult to program and design, seldom requested, extremely casual character (running out of room for them), appears on other platforms, based off of real person.

The last one is the real kicker for me. That means he may not give consent, and, even if he does, that means paperwork, compensation, and lots of legal issues. As a matter of fact, I'd rule it out entirely... but who knows? The concept of Kawashima in Smash has a tendency to stick in one's mind; if someone on the dev team thought of it, it may very well have grown.

Kawashima Want: 100%
SOOO UNIQUE. I do enjoy the Brain Age franchise and like the character, but I'm particularly attracted to what he could bring. The idea is just so... right.

Rosalina Prediction: 13.54%
Wow... the last two days have done wonders for her. Still, she shouldn't do too well.

Groose x4
Godot x1

Looks like the Groose is loose to end his day now!!

Yup. I'm not even that late; it's a Friday, so the Projected Update Schedule allots me more time.

DAY OVER​

​

MY BRAINAGE? Uh... low twenties, I think it was. ​

 

[Starcutter]

Godot x1

yes, please!

 

[BluePikmin11]

This was the most fantastic day in Rate Their Chances, I might take some of the information into consideration in my Kawashima thread.

 

[LoneKonWolf]

Looks like the Groose is loose to end his day now!!

(sorry the post is a little late, but here it is)

 

[Groose]

Dr. Kawashima (Brain Age)
5.15% chance
22.12% want

Some of you are probably wondering: how much did BluePikmin's rating inflate that score? Not that much, actually; without BluePikmin, he would have still got a 3.90% chance. As it is now, he's just a hundreth of a percent more likely than Rayman and slightly less wanted than the likes of Slime and Tingle. Doctor Kawashima apparently didn't understand how the brains of this game's players work.

Today we'll be rating the one and only Rosalina. Two days ago we received confirmation that she would be the fifth playable character in Mario 3D World; yesterday we received confirmation of an awesome Mario Galaxy-inspired stage with her nowhere to be seen. Will she be playable? Please rate Rosalina in chance and want today.

WARNING! CHALLENGER REAPPROACHING! Yeah, Waddle Dee is back. He posted the most nominations in a single day that I can remember; his 43 beat out the 37 I recorded for Lucina way back in the summer. But will he improve in his ratings? Please rate Waddle Dee in chance and want.

Tomorrow we'll be rating yet another of Nintendo's finest ladies. Please predict how Captrain Syrup will fare. Doing so can earn you extra nominations, like those won by Lonekonwolf... but wait! He hit it exactly on the dot, so he gets TEN extra nomiantions... and is the first person to perfectly predict twice, if I'm not mistaken.

AN APOLOGY: Sorry guys, I had planned on making a video to celebrate Day 100 and had even started doing so... but I don't have time to finish it or anything. I apologize for the lack of excitement... :/

 

[LoneKonWolf]

Tomorrow we'll be rating yet another of Nintendo's finest ladies. Please predict how Captrain Syrup will fare. Doing so can earn you extra nominations, like those won by Lonekonwolf... but wait! He hit it exactly on the dot, so he gets TEN extra nomiantions... and is the first person to perfectly predict twice, if I'm not mistaken.

like those won by Lonekonwolf... but wait! He hit it exactly on the dot, so he gets TEN extra nomiantions... and is the first person to perfectly predict twice, if I'm not mistaken.

WOOOOOOOOOHOOOOOOOOOOOOOOOOO!!!!!!!!!!!!!!!!!!!!!!
The second time I win the nomination's, I hit it on the dot!
that's also the amount I also rated Dr. K too I must be one lucky person