Slashdot Mirror
EA Simulation Correctly Picked Super Bowl Champs in September
Just_Say_Duhhh writes "Before the NFL Season started, the guys at EA Sports simulated the entire season using Madden 2011. The sim told them the Packers would win the Super Bowl. If only we had listened. What's even more interesting is that according to the article, they've picked the winner 6 of the last 7 years. Make that 7 out of 8!"
Clearly this means the NFL just throws the games based on the results of Madden!
.. accurate simulations are not news for nerds? I have no interest in American Football, or most other competitive sports, but I still think this is cool.
which is totally what she said
Maybe the NFL players are learning from Madden?
What else can happen when an unstoppable force collides with an immovable object?
Because according to *this* article, it picked the Steelers... http://blog.games.yahoo.com/blog/355-steelers-will-win-super-bowl-xlv-predicts-madden-11/ wtf?
But did it predict how bad the half-time show would be?
1. run enough independent simulations to predict each team as winning in one of them.
2. only report the right one
3. profit!
protip: replace "team" with "drug," and "winning" with "effective," for supermegaprofit!
"They were pure niggers." – Noam Chomsky
The super bowl sim predicted a steelers win they should have stayed with the pre season one. The prediction record is now 6 of 8
http://espn.go.com/espn/thelife/videogames/easims?id=6072052
What's even more interesting is that according to the article, they've picked the winner 6 of the last 7 years.
Does that mean the predicted the SB winner before the start of the season or before the start of the game?
If those were pre-season picks, I highly doubt they chose the Giants in 2007-08 or the Saints in 2009-10.
In other news, I still dislike EA games.
We should start a new Slashdot and return control to the geeks. It actually wouldn't be that hard to get some users to
let's say obama decides to flip a coin to start off the state of the union every year.
person A flips a madden coin a few times and makes note of it
person B does the same
person A's notes don't match up with Obama's flips
person A "it's flawed the madden coin is useless as a predictor"
person B's notes mostly match up with Obama's flips
person B "OMG what an awesomely calibrated simulation of Obama's flips we are running"
advice from slashdot?
Prove yourself: embezzle
You mean that american football is so boring that you can predict its result only by looking at statistics.
Nothing new here...
Same, maybe one day we can just not pay people but simulate the games... one can only wish.
PS: There is only one type of football, and it's only played with your feet. Don't worry, this is not just aimed at Americans, but my own countrymen/women too (Aussies). I suppose you can't call it the NHL (National Handball League) as that acronym is already taken.
This reminds me of a scam I heard about years ago.
1) Pick 1024 victims
2) Pick 1 volatile stock
3) To 512 victims, say the price will go up, to the others, say it will go down
4) Wait 5 days, then pick the half which were right
5) GOTO 2 (ok, you could use recursion instead, or a loop)
6) When you get to 8 victims, point out that you've been 100% accurate 6 times in a row and get them to invest a lot of money
7) Profit!
But they're reporting the right one before they know which one is correct. They're also doing it year after year.
So, do bookmakers use this for their odds calculations?
Where I grew up, their was an Italian delicatessen. They made great hoagies! It was also run by a couple of famlies, and they all drove Cadillacs. When you went in there, someone was always on the phone. Hmmm. When the racetrack nearby burned down and closed, they closed as well.
It was reopened by two guys who my parents knew. They said that the phone was constantly ringing from folks wanting to place bets.
But obviously, they made a tiny fortune on the betting business. So I have to wonder, how do bookies calculate their odds? Do they use IT technology? Or is it just a gut feeling? I'm not a betting man myself, but I don't mind other folks doing it.
And even if I did know that the delicatessen was a front for a bookmaker operation, I wouldn't have cared. As long as they kept making those hoagies. My tip: If you want to really experience a hoagie, find a mom and pop delicatessen in South Jersey.
Schroedinger's Brexit: The UK is both in and out of the EU at the same time!
Were you a graduate student at the time?
Caveat 1. Pray to $deity that the first 6 iteration's members don't know how to use the internet to junk your name, and the last 8 don't know how to use Google.
Mathnet (Square 1) covered this in 1990 with "The case of the swami scam" (in New York)
http://www.tv.com/mathnet/the-case-of-the-swami-scam/episode/236553/summary.html
If NFL football is this predictable, it's just further proof how idiotic and worthless the sport really is.
.. accurate simulations are not news for nerds? I have no interest in American Football, or most other competitive sports, but I still think this is cool.
What would really impress me is if they could predict the results of the Pakistani cricket team, including who was going to bribe who. The meta model for corruption would make this much more challenging than a game prediction.
Before the NFL season, an active NES ROM hacker simulated the entire season using Tecmo Super Bowl. The sim told him that a 7-9 team would win a home playoff game.
Gut feelings and odds have little to do with it. Bookies try to get enough people to bet on both teams. If too many people were betting on the Packers, they would move the spread so more people bet on the Steelers, and vice versa. They have to balance the money on both sides so they have enough to pay out. Most bookies take a percentage of the bet, or a "vig", so they get paid no matter which side wins. I ran a small operation in high school like this. The house always wins. If a bookie runs into the situation where his cash flow isn't as good as it should be, he usually lets them carry over the bet into the next week. That gets kinda hard with the Super Bowl. just my $0.02 (plus my percentage)
Fast, cheap, correct. You get to pick two.
ah yes, of course, because that paragon of journalism, Wired magazine, reported it.
this shit happens (sometimes unwittingly) to real scientists. somehow i'm just not trusting EA Games at their word...
"They were pure niggers." – Noam Chomsky
A Lot of what Bookies do to cover their odds isn't directly related to the event itself, but more to balance out their risk on what people are betting on. For example, If they took a bet on Team A vs Team B, and Team A was obviously much stronger, most people would want to bet on A, with very few betting on B. Based on that, they will adjust the contest either by giving the game a spread (For example, You only win the bet if Team A wins by at least 10 points) or giving higher odds to one team. (For example, Bet $1 and get back $5 if team B wins) These will both fluctuate before a game, as more people place their bets. The Bookie's goal isn't to correctly predict the game, but to end up in a situation where their risk is minimized, with roughly the same amount of money at to be paid out regardless of which team wins.
So I have to wonder, how do bookies calculate their odds?
According to an ex-bookie's clerk I know, mostly by watching other bookies very carefully and by trying to keep their overall liability to the punters to a minimum. Remember, the bookies are in this to make money, so they're actually looking to manage risk to themselves while maximizing the amount of trade they're doing; if they get it wrong, either the punters all go elsewhere (odds too low) or the punters all come running (odds too high).
"Little does he know, but there is no 'I' in 'Idiot'!"
Except, as you might notice, TFA is from September 2010, so it's not a retroactive reporting of the one simulation of many that happened to get it right.
A huge number of analysts picked the Packers to win the superbowl before the season started. You have to wonder if EA purposely tweaked the game stats to make sure that the packers would come out on top in their simulation.......
Monstar L
For the start of the book stats, more stats and even more stats.
If a team normally wins 1 in 10 games, offer the punter odds of 5/1.
Ultimately, after the book has run a week, see where the money is being placed. If the team receives a number of high value bets, reduce the odds to 3/1. And so on.
It's all pure maths. I wish I'd gotten into the game when I didn't need money to pay the bills.
This post contains benzene, nitrosamines, formaldehyde and hydrogen cyanide.
Perhaps this just means that American Football is unusually predictable.
What the story fails to mention is who they predicted would lose the Super Bowl. If they are truly on to something, the same method that predicts the Super Bowl winner should, also, be able to predict the winners of each conference. It says something about how accurately Madden Football reflects the relative strenghts of the teams, but it is only truly impressive if it can predict the Conference Chanpions with significant reliability.
The truth is that all men having power ought to be mistrusted. James Madison
Yes but would the story be appearing on slashdot if they had gotten it wrong? Or would slashdot have simply run with the story of a different simulator which happened to get it right?
With "7 out of 8" picks they may be on to something (depending on how many teams are actually viable and how many independent simulations are being run). But, then again, the fact that the simulation predicted Steelers would win the actual match, and that extrapolating backwards the system only becomes more chaotic, makes you wonder if they are just lucky.
When things get complex, multiply by the complex conjugate.
Perhaps this just means that American Football is unusually predictable.
Yet the injuries are not... This would seem to imply there is really no quantitative difference between 1st / 2nd / 3rd string players, despite widely held beliefs to the contrary. If player X had a season ending injury (or even just a game ending injury, early enough in the game) back in game 3 and player Y took over, it apparently doesn't matter much.
This would seem to be a big problem for the cult of personality that has grown up around individual players, so I suspect this story will have to be massively downplayed or treated as a statistical fluke.
"Science flies us to the moon. Religion flies us into buildings." - Victor Stenger
I should explain a bit more about the vigorish ("vig"). Rather than say to punters "We will take 10% of the bet" what the bookmaker does is arrange the odds so that overall a desired percentage less than what was paid in will be returned to "winners".
Maybe people seem to think the Packers will win, you offer them 2-5 odds [ 71% chance ], if they pay you $10 you'll give back $14 if the Packers win. Covering the other side who people seem to think are less likely, you offer better odds for the Steelers, 3-2 [ 40% chance ], if they pay you $10 you'll give back $25 if the Steelers win.
[ Notice that these chances don't add up - they're not true chances. Ensuring the advertised odds reflect chances of success that sum up to more than 100% is necessary to create the "vig" for the bookmaker. Check it out in any scenario where you're offered the chance to bet on outcomes ]
At those odds twenty people bet Packers, giving you $200, and ten people bet Steelers, another $100. $300 in the bank. If Packers win you pay out $280. If Steelers win you pay out $250. Either way you have a healthy profit. See?
Now, the problem with this is that someone who wants to bet on the Packers might meet someone who wants to bet on Steelers. They might agree mutually on odds in between the bookmakers odds, cutting out the vig. That's bad for the bookmaker, but good for them, (unless they lose the bet in which case their pockets are empty either way).
Once upon a time that was unlikely. But today there are thriving Internet sites for exactly this purpose. They take a straight percentage for making the arrangements and that's it. The percentage is smaller than the bookmaker's vig and is trasparent - everybody knows they're paying it.
1. Sum up net capital of club investors
2. Sort descending.
3. First entry in list wins.
Slashdot social media options: AIM, ICQ, Yahoo, Jabber and Mobile Text. Why no MySpace?
why don't we hire them to do a climate change model?
Gut feelings and odds have little to do with it. Bookies try to get enough people to bet on both teams. If too many people were betting on the Packers, they would move the spread so more people bet on the Steelers, and vice versa. They have to balance the money on both sides so they have enough to pay out.
Come on, this is slashdot here. Try for a CS or automotive analogy.
http://en.wikipedia.org/wiki/Binary_tree
Basically a bookie builds a binary tree where the top level node (aka the bookie) has two child nodes, folks betting on team A and folks betting on team B and the bookie rebalances the tree by screwing around with the odds until the top level of the tree has about the same number of $ on its two child nodes, more or less. A successful bookie has a lot of customers first to balance the tree and secondly because he skims off tips in a microtransactional approach, more or less.
Now if you want the standard slashdot car analogy, bookies change odds like inflating and deflating the tires on the left and right side until the car tracks more or less straight without pulling either side, sorta.
"Science flies us to the moon. Religion flies us into buildings." - Victor Stenger
Yes, but they work at Chase, Discover, Citi, and other major credit card issuers. They also work at Transunion, Experian, and Equifax in developing new credit reports. Instead of phones, they use SAS Software, together with direct mailings and commercials.
This kind of modeling has been going on for decades. Heck, sometimes it is beginning to look like psychohistory with its accuracy.
Even in sports, this application is nothing new. Baseball is, to my knowledge, where statistics (economics if you like) was used first. The approach is called sabermetrics. Baseball was the ideal candidate because of the number of games played over a season. With so many more observations than other sports, the likelihood of a correct analysis rises.
Hoist Number One and Number Six.
From the article:
"September 3, 2010"
They reported one result, well in advance. On the other hand...
"Although the Packers are slated for Super Bowl glory, their 12-4 regular-season record will be second-best overall to the Indianapolis Colts and Baltimore Ravens, who’ll go 13-3, albeit in a different conference. The Packers will tie Favre’s Minnesota Vikings for the NFC’s top record but will be the top seed in the conference due to a stronger record against divisional foes."
Yes, most of their predictions were wrong. :)
And put $50 on "Little Louie" to win.
Gamingmuseum.com: Give your 3D accelerator a rest.
Why laud EA Sports, engineers of a no-competition contract with the NFL, whereby nobody else can make an NFL game because they hold an exclusive license? For a community that hates all things closed and proprietary, EA is the MSFT of video games.
I think that octopus Paul has a better record.
If he mentions the pig again, use two hammers!
(1.21 gigawatts) / (88 miles per hour) = 30 757 874 newtons
No they calculate their lines so that they get equal amounts of money betting on both sides. They make their money on the "edge", a precentage, typically 10%, of the bet.
Absofrigginlutely!
Wired reported EA's prediction in September, they were proven right yesterday. Not sure where the trust factor comes in here.
Looks like the graphics are finally good enough, we thought we were watching a live game, we were just watching two kids playing Madden....
I, for one, welcome our new robotic overlords.
This is like all the "momtuition" crap on CNN. CNN only reports the successful cases of mom using her mom-sense to "know" that the doctors are wrong and by luck is. They don't report the bulk of the cases where mom is wrong (with possibly fatal results) and the doctors were right.
As a non-American, baseball seems to be a combination of a tedious stock market show churning out reams of statistics, interspersed occasionally with a bit of rounders (a game popular in the UK with children).
To have a right to do a thing is not at all the same as to be right in doing it
http://en.wikipedia.org/wiki/Parimutuel
Doesn't baseball also have a lot fewer injuries than football? That would also make it a lot easier to predict.
There is no -1 Disagree mod. Slashdot.org/faq defines mod options. USE IT.
other response are technically correct, but from a theoretical perspective, the bookie (less bid/ask spread) is computing a risk neutral price based on actual betting behavior. the optimal odds are the same ratio as the # of bets placed on each side. this guarantees the bookie doesn't lose money. add some extra fee for placing the bet and you always make money. the original odds might be set based on some empirical estimate, but the final odds used are always based on ratios of bets placed.
BSD is for people who love UNIX. Linux is for those who hate Microsoft.
On paper or in a simulation, the Packers definitely had what was needed to win a Super Bowl. The NFC competition was pretty shallow and the Packers have great players on both defense and offense. What made last night's victory so unlikely was all the injuries they dealt with. To have that final defensive stand without Charles Woodson on the field was amazing, to still pass the ball 60%+ of the time without Donald Driver was equally impressive. And they had many more injuries. The Packers were complete underdogs but managed to pull through. Take a Madden simulation with their injured players pulled from the game and it's very doubtful they would beat the Steelers.
So this is a case where the simulation picked the right team but it's more of a coincidence than a super-accurate prediction.
"From the depths of my skeptical and rationalist soul, I ask the Lord to protect me from California touchie-feeliedom."
Ah. But you are leaving out the important part. Beer.
Hoist Number One and Number Six.
Yeah everyone knows Wired has a time machine. They can't be trusted.
There is only one type of football, and it's only played with your feet.
The term football was originally used to describe a number of different games that were played on foot.
Someone flopped a steamer in the gene pool.
Historical correction: The word football was originally used to describe a game that eventually became Association Football (the kind you play with your feet), and Rugby Football. The latter was brought to the US, and when poorly executed lead to a lot of deaths. So the rules were adjusted (and influences from other games also entered play), and American Football was born. Also, the word "soccer" comes from "association" and was in common use in the UK prior to WW2. You can watch old matches from the 30s and hear commentators call it "soccer." I'm also told that some Kiwis call Rugby 'football' and call football 'soccer.'
There is more to science than physics!
www.iomalfunction.blogspot.com
Well, if you actually bother to read TFA - you'll note the date is prior to the start of the season. So you're scenario is basically impossible.
Actually its quite simple: While the big name plays don't affect the overall outcome of the game what they DO affect is how many butts are in the seats, and since it all comes down to making $$$ that is why they get the big bucks, because they draw more people and thus make more money for the owners. See? Simple.
ACs don't waste your time replying, your posts are never seen by me.
I notice the resounding silence on the right over the victory of the only publicly-owned non-profit NFL team over all of the privately owned for-profit teams.
In the Faux News/Glenn Beckistan universe this cannot ever, ever, ever happen.
OK, that's one for one. Whoop-dee-doo. What makes this interesting is the 7/8 claim. Please produce seven years of historical articles. Maybe they're out there and all vetted; I don't give a fuck.
"They were pure niggers." – Noam Chomsky
ERROR: All columns specified in a GROUP BY query must use aggregate functions.
Local music(to upstate NY). http://gnarfel.com/ radio.
The house always wins.
Uhh not quite with sportsbooks. In fact, this makes 2 of the last 4 Super Bowls where most books lost money. Don't believe me? ESPN's Vegas Insider Chad Millman is reporting that Vegas lost money this year. The problem with your logic is you assume that the linemakers are making correct lines in anticipating the action. However, just like the Giants/Pats SB, Vegas made the line so "soft" they couldn't adjust the line to even out the wagering. Your reasoning that they will just change the line as the money flows in is fatally flawed because with big money events like the Super Bowl, creating a big "middle" is death to Vegas and the Sharps will hammer both sides of the lines and potentially break the book. This year, there was too much money on the Packers and the Over and once the 2pt conversion was successful, Vegas was done for.
All week long bookmakers such as the LV Hilton were reporting that they didn't want to move the line from 2.5 to 3, but were forced to and then had to play with the juice (-130 for +3 Steelers) and that there was no way they were ever going to 3.5 to create the middle. And as for 2007, you can thank all that New York money that flowed into Vegas at +450 Giants Money Line for causing the red numbers in 2007.
No. Bookies use what people are betting on to calculate odds. That's why they are constantly changing. If some rich guy puts 100 grand on a team, the lines will shift to accommodate his bet (to encourage people to bet on the other team). In places where sports betting is legal (like Mexico and Vegas) the second half betting lines sometimes move very rapidly during half time, right before they shut down because everyone is betting one team. This is really common (and a good bet!) when a team that is obviously better but had a bad first half. For example, Alabama played the razorbacks back in October. They were down 7 at the half and the second half spread was 7 points. Pretty safe to assume that Bama was going to win that game and cover the second half spread. Which they did. I managed to get in my bet seconds before the bookie closed the lines on that half because everyone was doing the same thing.
A bookie wants the betting to always be even for both teams because then he is hedged and can't lose money. He just takes his 10% or whatever off the top (well really out of the middle). If the bookie is unable to drum up enough betting for one team he will simply close the lines.
oh, the "What? I was wrong? Well, let's shift the argument a few feet over to the left and try again.. I'm not wrong yet, now."
# (/.);;
- : float -> float -> float =
"There is only one type of football"
You mean the boring dreadful soccer I suppose. Aussie rules is a far more excting and interesting game than Soccer.
And I grew up a soccer fan in my early years. When I started to watch Aussie rules, soccer just became a lame boring waste of space to me.
But if a new player came in and was just as good, would that new player be the star? I believe that most sports viewership is for tribalist reasons, otherwise people wouldn't watch teams for decades when the old players have long since retired.
Neutrals may watch a game to see a star player, but is that really a significant proportion of the market?
Hi DRSquare! In this case while what you are saying appears to be true, if you look at it logically, you are forgetting something about us hairless monkeys that makes all the difference here...personality sells and what you need is for the athlete to be both good at the game AND a personality to put butts in the seats.
This is why players like Elway and Montana are remembered long after they are out of the game, because they had personality which made them more than great athletes it made them entertaining and that is ultimately what is it about, no different than pro wrestling although I'll get labeled hater just for daring to say it.
So while the games may not be rigged (well most of the time, after reading "North Dallas Forty" I'm not so sure on some of them) winning will only get so many butts in the seats if the players have the personalities of rocks, you need a draw, someone the fans can root for. Be it the plucky kid that is green but has a good heart, the grizzled veteran trying to have that one great season before hanging up his cleats, or the rock star with the million dollar arm that throws like there is just no tomorrow people want to root for the hero. And THAT is what adds up to a serious difference in profits, especially with items like merchandise. Hell look at how many bought 49ers stuff with Montana was king, or Dallas gear when they had Staubach/Dorsett? That is serious money for the owners, which equals serious money for the player that grabs them headlines.
ACs don't waste your time replying, your posts are never seen by me.