Baltimore Ravens 16 Game Season Simulation
by Bruce Raffel on Sep 9, 2010 6:00 AM EDT
in Xfinity Sponsored Posts
Email
PrintI received an email form Jimmy Shaprio about a website called www.PredictionMachine.com. It comes to us courtesy of Paul Bessire, who is a statistician and sports writer and has developed a simulator that predicts the outcomes of games by playing them over and over for 50,000 times until he gets an average score. This means his "Predictalator" has actually already played the entire 2010 NFL season 50,000 times!
Unfortunately, his Prediction Machine has the Jets not only winning the Super Bowl, but beating the Baltimore Ravens in the season opener on Monday night. At least he has the Ravens finishing atop the AFC North with an 11-5 record.
Check out his email to me after the 'Jump' with a link to his site and some of the prediction categories.
I thought you might be interested to see the Ravens 16 game simulation courtesy of Paul Bessire of www.PredictionMachine.com. If you’re interested in speaking to Paul about these numbers via phone or email questions, please let me know. Or you’re more than welcome to just post the numbers yourself as long as you credit PredictionMachine.com.
PredictionMachine.com strives to be the most accurate and trusted source for predicting sports outcomes. The state-of-the-art, Prediction Machine, was built by Paul Bessire, a statistician and veteran sports writer, who offers his in-depth analysis and unique takes through his blog posts and game predictions.
In sports, matchups dictate outcomes. Instead of data mining to attempt to find historical trends that may lend some insight into future games - a common, yet misleading approach for many other predictive models - PredictionMachine.com's accuracy stems from the Predictalator (a fancy name for the Machine), which plays an upcoming game 50,000 times before it's actually played. It’s actually played the whole NFL season 50,000 times already.
The results are always fun to look at. Please see the links below plus a season simulation for the Ravens.
Predicted Standings: http://predictionmachine.com/
Divisional Previews (which includes a Super Bowl prediction): http://predictionmachine.com/
Over/Under Win Totals : http://predictionmachine.com/
Team Stats: http://predictionmachine.com/
Fantasy Projections: http://predictionmachine.com/
|
Baltimore Ravens Schedule |
||||
|
Week |
Opponent |
Ravens Score |
Opponent Score |
Win % |
|
1 |
@ Jets |
14.1 |
18.1 |
38.3 |
|
2 |
@ Bengals |
19.4 |
18.4 |
52.8 |
|
3 |
33.7 |
12.7 |
86.2 |
|
|
4 |
@ Steelers |
19.9 |
17.0 |
58.6 |
|
5 |
28.0 |
14.4 |
78.8 |
|
|
6 |
@ Patriots |
23.0 |
24.9 |
45.1 |
|
7 |
24.5 |
11.6 |
79.6 |
|
|
8 |
Bye |
|
|
|
|
9 |
30.7 |
17.2 |
77.1 |
|
|
10 |
@ Falcons |
23.3 |
22.4 |
52.5 |
|
11 |
@ Panthers |
23.3 |
19.0 |
61.5 |
|
12 |
29.6 |
11.2 |
85.2 |
|
|
13 |
Steelers |
23.6 |
14.6 |
72.1 |
|
14 |
@ Texans |
26.3 |
23.3 |
58.0 |
|
15 |
29.5 |
22.1 |
65.7 |
|
|
16 |
@ Browns |
29.9 |
14.9 |
88.1 |
|
17 |
Bengals |
22.7 |
15.7 |
68.0 |
|
|
|
Points/Game |
Point/Game |
Total Wins |
|
|
|
25.1 |
17.3 |
10.7 |
About Paul Bessire Paul Bessire, who built the Predictalator, is one of the foremost authorities on mathematically modeling and analyzing all sports. For more than six years since earning his Master's degree in Quantitative Analysis (with a focus on Finance and Sports), he has turned what was his longtime hobby of predicting and writing about sports outcomes into his full-time profession. About PredictionMachine.com Launching during the 2010 Super Bowl, PredictionMachine.com strives to be the most accurate and trusted source for predicting sports outcomes. The state-of-the-art, unbiased Predictalator ("the Prediction Machine") was built by Paul Bessire, a statistician and veteran sports writer, who offers his in-depth analysis and unique takes through his blog posts and game predictions.
Over that time period, Paul has had great success, like accurately choosing the winner of: five of the last six Super Bowls, including last year's four-point (exactly) win by Pittsburgh; five of the last six World Series, including the Yankees' 2009 win in six and the 2006 Cardinals win in five (both to the game); five of the last six Stanley Cup champions; and five of the last six NCAA Tournament champions (as of Selection Sunday).
Paul also just bested more than 100 members of the media to win the 2009 Richard GardnerCharity Pro Football Pick'em contest in its first year in existence.
Most recently, Paul Bessire served as the Product Manager of Quantitative Analysis and Content for FoxSports Interactive. There, Paul was responsible for the math behind the historical simulation sports games at WhatIfSports.com as well as any FoxSports.com, Scout.com or WhatIfSports.com articles that were created from that technology. In that role, Paul wrote over 1,000 articles and conducted more than 500 radio interviews. He has also collaborated on countless other print, online and TV features with many of the most recognized sports media (more info here), some sports leagues and teams and many of the best minds in the sports analysis business.
PredictionMachine.com is Paul Bessire's latest endeavor, which features the Predictalator – the most in-depth, state-of-the-art sports prediction software ever created.
Paul was also recently the featured speaker at the annual "Symposium on Statistics and Operations Research in Sports"that was held in St. Louis around the 2009 MLB All-Star Game.
In sports, matchups dictate outcomes. Instead of data mining to attempt to find historical trends that may lend some insight into future games - a common, yet misleading approach for many other predictive models - PredictionMachine.com's accuracy stems from the Predictalator, which plays an upcoming game 50,000 times before it's actually played.
This is the best way to account for every possible interaction between players (playing or not playing/injured), coaches, officials, fans (homefield advantage), weather and more. Any individual who tries to make a prediction, coaching decision, etc. is attempting to do this type of comparison in his/her head. Without the aid of a "machine," these comparisons are very difficult to do quickly and without personal, subjective biases. The Predictalator can handle it – and better than any other application available.
There is an answer for every game, event or contest and PredictionMachine.com will be your place to get it. At PredictionMachine.com, we play the game 50,000 times before it's actually played!
Everything in sports comes back to the numbers. Don't fight the numbers. Trust the numbers.
15 comments
|
0 recs |
Do you like this story?
Comments
Another source for football
I don’t like the outcome of the Prediction Machine but there is some good information on that site.
There is some favoritism there
It looks like he just assigned a number to each team for how good they think they are and went with that. That Flacco will only have 3600 yards and we will only rush for under 2000 is just silly. From what I have gathered, they expect us to actually have a collapse on offense in every portion instead of getting better at all.
Just because of that…. I’m not buying into this one.
Agreed
If Flacco were injured (he’s not) or if our WR situation was a mess (it’s not), then I might believe the 3600 passing yards.
This is where I think his model breaks down in that it may not take historical trends into account. For example, all other things being equal in terms of injuries and receivers, historically QB’s in their third full year increase their statistical output in a positive way that is statistically significant -this positive increase is easily within three deviations of the mean for all outcomes.
And I agree with the poster above about injuries. I think you would almost have to run this model each week. Anything more than a week out, because of unforeseen injuries, is just throwing darts at a board…
I hope they put yanda at right tackle and Chester at guard. I don’t wanna take a chance on cousins.
by Raven_all_day on Sep 9, 2010 11:33 AM EDT via mobile up reply actions
The problem with this kind of analysis
is several-fold. I think it would only work really well for an established team (+3 years with the same players at the same positions). With that scenario, you have established norms (and records) to take statistics from. In the case of the Ravens, Almost our entire offense has less than 3 years being on the same field – hell, for some, less than 3 weeks. How are you going to model what players who’ve virtually never played together before will do?
Also, if, as implied above, there is a bit of favoritism or bias or whatever in assigning a base “value” to the team, well, you can guess how that affects outcome.
The third thing is why play the season 50K times? In reality, these games are played once, with all sorts of variable factors (weather, injuries etc.) affecting the outcomes. While I’m not a statistician (and really only understand the basics), by the time you’ve sim’d the season 50K times, with little or no account taken for the variables, I would say you have no more chance of an accurate prediction than you would if you just flipped a coin once.
Somebody sim the season on Madden and you’ll probably get as good a prediction.
Just my opinion. YMMV.
That and the NE prediction
We destroyed them in the playoffs… we didn’t just beat them, but DESTROYED them. How is it that they beat us? They have made very few offseason moves, especially at their weaknesses, while we have added nothing but ProBowl players on offense and have added very qualified players on defense.
It just screams like every year of people saying NE, and SD will be the best teams in the AFC. Every year, I have to hear this is the Charger’s year! It hasn’t happened when you had all your good players, it certainly isn’t going to happen with 2 of them sitting out the year.
by Mstevens_Design on Sep 9, 2010 3:06 PM EDT up reply actions
Mark Sanchez...
…doesn’t scare anyone. In my opinion, the Ravens can win games in as many ways as the Jets can lose them. Cam has a lot of weapons and a pretty good QB. The D is better than it was last year at most positions and while Ed Reed being injured is a major loss, the fact that Frank Walker won’t be on the field is a major plus.
Add in a more experienced Flacco and Rice, and new skill position players and you have one of the better teams in the league.
As for the Jets, I respect their team but they won’t have Santonio on Monday and they will have a 2nd year QB all season. I’m not worried.
Fortunately each game isn’t played 50,000 times. It’s played once. We’ll find a way into that 38% window on Monday night (although I contest his calculus on these simulations)
"Cam, we're not going to have any issues because we're going to be winning."
--Joe Flacco
"Just win baby. Yeah, I stole that."
--Jon Gruden
although I contest his calculus on these simulations
Agreed. There’s absolutely no way to know whether his method holds any water without comparing several years worth of projections to the actual season. Since this is the first full season it will be in use, there’s no reason beyond entertainment value to pay any attention to it. I’d be interested to check back in on it in 2015.
Water covers 2/3 of the Earth's surface. Ed Reed covers the rest.
“Good news everyone, I developed a prediction machine ..”
"wherever Brad St. Louis is and Shayne Graham is about to be." -R.F. Mehl
So that's what it would be like if I had invented that prediction machine
Well at least I still have the finglonger.
by Roa on Sep 9, 2010 8:21 PM EDT up reply actions
Comments For This Post Are Closed
Secondary Sidebar
User Tools
Related FanPosts on Baltimore Beat Down
FanPosts
Community blog posts and discussion.
Certain photos copyright © 2012 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
