Game Probabilities - Week 5
Weekly game probabilities are available now at the nytimes.com Fifth Down.This week I also lead in with my take on how much luck can have to do with game outcomes.
Best Games of the Decade
Check out the Win Probability graphs and play-by-play of your favorite team's biggest comebacks and most exciting games since 2000. An explanation can be found here. Just select a year, a team, or 'any', and start clicking:
Or search for all the games for your favorite team:
Oct 8, 2009
Subscribe to:
Post Comments (Atom)
Posts
26 comments:
Win Chance GAME Win Chance
0.29 Cleveland at Buffalo 0.71
0.64 Dallas at Kansas City 0.36
0.65 Minnesota at St. Louis 0.35
0.11 Oakland at NY Giants 0.89
0.13 Tampa Bay at Philadelphia 0.87
0.73 Pittsburgh at Detroit 0.27
0.47 Washington at Carolina 0.53
0.34 Cincinnati at Baltimore 0.66
0.37 Atlanta at San Francisco 0.63
0.65 Jacksonville at Seattle 0.35
0.37 Houston at Arizona 0.63
0.26 New England at Denver 0.74
0.71 Indianapolis at Tennessee 0.29
0.57 NY Jets at Miami 0.43
Brian explained how his predictions work in this post.
No calculation faults this week?
Anybody mind if I post my College Football picks?
I'm sure no one will.
do you use this same method?
This is a similar methodology.
Well, here goes. These are my preliminary PICKS 10/7/2009.
NEBRASKA -3 1/2
LOUISIANA TECH +10 1/2
BOSTON COLLEGE +13 1/2
MINNESOTA -3 1/2
AUBURN -2 1/2
BALL STATE +13 1/2
GEORGIA +1 1/2
MISS STATE -2
MARSHALL -4
VIRGINIA - 6 1/2
ALABAMA -5
CONNETICUTT +8
NC STATE -14
MARYLAND +13 1/2
TOLEDO -7 1/2
STANFORD PICK 'EM
IDAHO +4
GEORGIA TECH +3
MICHIGAN +8
MIAMI-OHIO +19 1/2
THAT'S 20 PICKS,,,I AM AVERAGING A 77.8% ATS CORRECT RATE FOR THE PAST 3 WEEKS.
SO, 15-16 SHOULD BE DEAD ON.
LET'S SEE HOW I DO.
JMED2 in Georgia
jim from GA, 78% ATS? Thats pretty ridiculous over a 3 week span. Can you post last weeks picks from NCAA?
I meant ridiculous in a positive not, not in a demeaning way.
what do you show the win probability of the giants at if eli doesn't play? we all know carr sucks.
JMD2,
What methodology do you use to formulate your college football picks (statistical, hunch, etc.)?
I use strength of schedule, both who you have played and will play, home and away, history of the Money Line and Spread movement...etc, etc.
JMED2 in GA
Not really the same...at all.
Brian: What do you mean when you say in your NYT post that "When we statistically compare the real N.F.L. to coin flips, only about half of all games are determined by the relative strength of each opponent?" Clearly, far more than half of games are won by the stronger team. Are you saying that half the variance reflects skill, and half is luck?
Guy: Just because a game was won by the "stronger" team, doesn't mean that that stronger team won due to skill. In other words, even strong teams get lucky.
from my understanding of the luck articles, 50% of the game is determined by skill and 50% is determined by luck (random events) therefore, a stronger team should win on average 75% of the time [50% skill + 25% luck (50% luck / 2 teams)]. Is that a correct understanding?
That's correct.
Guy-That's a rough approximation based on a couple things--1) a comparison of a Monte Carlo model of the NFL season using 50/50 coin flips and the actual distribution of season win totals; and 2)The (rough) ratio of variance of the binomial distribution to the variance of the actual NFL above and beyond the binomial variance.
Brian, thanks. When I've looked at this, I get about 44% of the variance from binomial variance and 56% team strength -- sounds similar.
Out of curiosity, how often does the betting favorite win in the NFL? (Not beat the spread, just win the game.) Is it around 75%? Lower?
If I recall I saw 52 and 48, but with a different method.
It varies from year to year, with the average being just under 70%.
College Picks: 12-8
60%,,,Personally disappointed.
Shockers: Auburn, Georgia, Miss St. & Stanford LOSSES.
JMED2 in Georgia
Great job, as always
man whoever posted the college picks needs to work on that system. 60%?? Not much better than a blind monkey picking on paper.
College Shockers: Auburn, Georgia, Miss St. & Stanford LOSSES. That's the difference between 16-4 (80%) and 12-8 (60%).
Win Chance GAME Win Chance
0.29 Cleveland at Buffalo 0.71.........LOSS
0.64 Dallas at Kansas City 0.36 .......LOSS
0.65 Minnesota at St. Louis 0.35.......WIN
0.11 Oakland at NY Giants 0.89 ........WIN
0.13 Tampa Bay at Philadelphia 0.87....WIN
0.73 Pittsburgh at Detroit 0.27........LOSS
0.47 Washington at Carolina 0.53.......LOSS
0.34 Cincinnati at Baltimore 0.66......LOSS
0.37 Atlanta at San Francisco 0.63.....LOSS
0.65 Jacksonville at Seattle 0.35......LOSS
0.37 Houston at Arizona 0.63...........WIN
0.26 New England at Denver 0.74........WIN
0.71 Indianapolis at Tennessee 0.29....WIN
0.57 NY Jets at Miami 0.43.............PENDING
AGAINST THE SPREAD: 6-7 W/ONE GAME PENDING
what do you mean against the spread? brian's not giving probabilities of a team covering the spread, he's giving the probabilities of teams winning the game outright.
straight up the results are:
Win Chance GAME Win Chance
0.29 Cleveland at Buffalo 0.71.........LOSS
0.64 Dallas at Kansas City 0.36 .......WIN
0.65 Minnesota at St. Louis 0.35.......WIN
0.11 Oakland at NY Giants 0.89 ........WIN
0.13 Tampa Bay at Philadelphia 0.87....WIN
0.73 Pittsburgh at Detroit 0.27........WIN
0.47 Washington at Carolina 0.53.......WIN
0.34 Cincinnati at Baltimore 0.66......LOSS
0.37 Atlanta at San Francisco 0.63.....LOSS
0.65 Jacksonville at Seattle 0.35......LOSS
0.37 Houston at Arizona 0.63...........WIN
0.26 New England at Denver 0.74........WIN
0.71 Indianapolis at Tennessee 0.29....WIN
0.57 NY Jets at Miami 0.43.............PENDING
so far this week 9-4
i agree it is easier to pick the winner straight up than ATS, but the way these probs can be used in betting is to multiply the probs by the bookie odds to find edges. if the result is greater than 1.0 than you are getting favourable odds.
e.g. for tonight ny @ mia
pinnacle pays 1.69 for ny, 2.33 for mia
for ny 0.57 x 1.69 = 0.963
for mia 0.43 x 2.33 = 1.002
in this case the pinnacle odds are accurate in relation to the probabilities, so there is no edge on either side.
TheEdge,
Can you do this one more time so I understand you correctly? You lost me when you get the numbers 0.963 and 1.002. How and what are you comparing those against to see if there is an edge or not?
Ok for all you have that have posted Auburn, Georgia, Stanford and Miss st as "shockers" all for of those games had a line no greater than 3. None of those games would qualify as a shocker! The only shocking part is how they got beat, not that they got beat. A "shocker" would be Duke beating a 15 point favorite like NC State.
Post a Comment