New Twist On NFL Predictions

MotorCityMadman · Sep 23, 2014

The Microsoft Engine That Nailed The World Cup Is Predicting Every NFL Game — Here Are Its Picks For Week 4

REUTERS/Mike BlakeMicrosoft's Cortana went 10-6 in its NFL picks in Week 3.
Cortana, the Windows phone virtual assistant that's using a Bing Predicts algorithm to predict every NFL game, is now 29-19 on the year. It's a good record but not a great one.
Las Vegas favorites are 27-21 overall (not against the spread).
Nate Silver's ELO model is now 31-17. It's beating Cortana after correctly picking 12 of 16 Week 3 games.
Cortana uses all the data you'd expect to go into one of these prediction models — home-away effect, historical data, and offensive and defensive advanced stats — but it also uses a public sentiment variable that pulls in Facebook and Twitter data.
Cortana and Vegas differed on two games in Week 2, and each won one of them. Cortana wrongly picked the Browns to beat the Ravens. Vegas wrongly had the Bills as the favorite over the Chargers.
Here are Cortana's picks for Week 4 (Vegas betting lines as of Tuesday afternoon are in parentheses):

New York Giants vs. Washington Redskins — Redskins win, 64.4% chance (Redskins -3.5)
Green Bay Packers vs. Chicago Bears — Bears win, 58.2% chance (Packers -1.5)
Buffalo Bills vs. Houston Texans — Bills win, 53.3% chance (Texans -3)
Tennessee Titans vs. Indianapolis Colts — Colts win, 68.6% chance (Colts -7.5)
Carolina Panthers vs. Baltimore Ravens — Ravens win, 65.9% chance (Ravens -3)
Detroit Lions vs. New York Jets — Lions win, >50% chance (Lions -1.5)
Tampa Bay Buccaneers vs. Pittsburgh Steelers — Steelers win, 78.9% chance (Steelers -7.5)
Miami Dolphins vs. Oakland Raiders — Dolphins win, 58.2% chance (Dolphins -4)
Jacksonville Jaguars vs. San Diego Chargers — Chargers win, 83% chance (Chargers -13)
Atlanta Falcons vs. Minnesota Vikings — Falcons win, 56.6% chance (Falcons -3)
Philadelphia Eagles vs. San Francisco 49ers — 49ers win, 58.2% chance (49ers -5.5)
New Orleans Saints vs. Dallas Cowboys — Cowboys win, 53.3% chance (Saints -3)
New England Patriots vs. Kansas City Chiefs — Patriots win, 67.4% chance (Patriots -3.5)

Cortana and Vegas disagree on three games: Saints-Cowboys, Bills-Texans, and Bears-Packers.
Here are Cortana's results so far this year:

Week 1 — 10-6
Week 2 — 9-7
Week 3 – 10-6
Total — 29-19

Check out how Cortana makes these picks here >

http://finance.yahoo.com/news/microsoft-engine-nailed-world-cup-204019983.html

Not bad +10 units minus juice.

DrMove · Sep 23, 2014

cheersgif

RookieDL · Sep 23, 2014

I don't believe the Cortana system is taking into account the odds. It's just a pick'em so the 10units comment isn't accurate. Correct me if I'm wrong. Thanks for posting, haven't heard about this algorithm system.

MotorCityMadman · Sep 23, 2014

RookieDL said:
I don't believe the Cortana system is taking into account the odds. It's just a pick'em so the 10units comment isn't accurate. Correct me if I'm wrong. Thanks for posting, haven't heard about this algorithm system.

Looks like the spread is taken into account. Spread is there.

wizard1183 · Sep 23, 2014

I agree with the cowboys. They lost 49-17 in the NO dome last yr and now at home vs a saints team who IMO is not that great this yr. they've been the fav all 4 weeks and and have lost the first 2 away and looked like shit at home vs Vikings even tho they did cover. I feel cowboys on this revenge game as a home dog wins outright.

The stat as a team who played 2 away games then a home game and is now away (NO) is 44-79 35.8% SU and 52-65-6 44.4% ATS winning last 3 in a row. If I throw in that they were favs all 3 wks and now a fav, they're 2-1 winning their last.

I realize cowboys aren't that great, but saints shouldn't be a fav when cowboys are 2-1 SU and ATS and they're at home. Saints are just 1-2. In no way do I say slam the cowboys, but my intuition and feel of the game is leaning cowboys. I'd love input on this...

TimesYours · Sep 23, 2014

I would disagree with Dallas and the Jets on that list. My team, the Eagles, just seem to have no pass rush this year whatsoever! None, zilch,zip! Add in the fact that Nate Allen and Brandan Fletcher can't cover a Pop Warner player we're gonna be scored upon. The only way it seems so far that we win is by outscoring our opponet. With 3 starting O lineman out of the line-up this week we;re due for a let down.
So far we've been, I won't say lucky but fortunate to be 3-0.

MotorCityMadman · Sep 23, 2014

wizard1183 said:
I agree with the cowboys. They lost 49-17 in the NO dome last yr and now at home vs a saints team who IMO is not that great this yr. they've been the fav all 4 weeks and and have lost the first 2 away and looked like shit at home vs Vikings even tho they did cover. I feel cowboys on this revenge game as a home dog wins outright.

The stat as a team who played 2 away games then a home game and is now away (NO) is 44-79 35.8% SU and 52-65-6 44.4% ATS winning last 3 in a row. If I throw in that they were favs all 3 wks and now a fav, they're 2-1 winning their last.

I realize cowboys aren't that great, but saints shouldn't be a fav when cowboys are 2-1 SU and ATS and they're at home. Saints are just 1-2. In no way do I say slam the cowboys, but my intuition and feel of the game is leaning cowboys. I'd love input on this...

I always love a home team getting points. Saints are no bargain to be favored on the road. As long as the Cowboys don't beat themselves they should be alright,
Coming back from 21 points down should give them a lot of mo.

Green17 · Sep 23, 2014

Cowboys tied a franchise record with that comeback win against StL. Might make their heads a bit bigger coming home. Tricky spot for Dallas I think, even at night.

DT32 · Sep 24, 2014

Not ATS..

For comparison, Las Vegas favorites are 17-15 overall and Nate Silver's ELO model is 19-13. Considering that the engine is just predicting winners and not results against the spread, 19-13 is just an OK record.

Read more: http://www.businessinsider.com/microsoft-cortana-nfl-picks-week-3-2014-9#ixzz3EFrlmwAf