I think basically what I was trying to say is that chaotic systems aren't practical to model statistically on short timescales, and short timescales is relative to the system itself.<snip>

It's like trying to predict the weather, even with huge super computers hell bent on figuring this stuff out we can at best predict trends, then you get your Katrina's

And I would say that weather predictions on a short time scale, say up to 12 hours, are fairly accurate and accuracy improves as the time scale decreases.

With respect to NFL games, I suspect that on short time scales (say the last 30 minutes of game time), statistics from the first 30 minutes are fairly good at predicting the outcome, but perhaps not as good as weather predictions. It is with long time scales, say 5 to 10 years, that game outcome predictions between various teams fall apart.

John

 

jpan, only course weather predictions are accurate. I live near lake Ontario, the dynamics here are complex the weather reports are NEVER accurate when there's rain or snow concerned, there's too much banding and even a temperature difference of a few degress can result in a light dusting of snow or a foot, and on the drive home we'll pass through two or three spots that are alternately wet/dry from the rain banding.

Jpan the in game statics are useless, only the worst of hardcore gamblers will bet period to period, a whole game is a basic start with over under points. One single bad play with an ankle or knee injury and that game is ruined statistically, and the whole season for that one team is thrown up in the air not to mention the repercussions to the rest of the teams they have to play through the remainder of the year.

Prediction is impossible. Rough guesses, and fairly accurate are not predictions. On a global scale hurricane tracking is often off by hundreds of miles in a single day, two days out it could be twice that. Or in the case of tornado dozens of mile in hours.


The sheer scale of things make it impossible. When you say fairly accurate that's only on a human scale. The amount of energy involved and the mass upon which it's acting on is really beyond human comprehension.

Even 12 hour predictions aren't really predictions, they can't predict gust duration direction and amplitude even in course areas local system instabilities are too dramatic. They can predict general rainfall over a wide area, but not peaks and dips for specific locations.

 

Well, on weather, I think we will disagree. Cleveland is also near a Great Lake, and I find those aspects of weather that I expect to be predicted are predicted quite well. By that I mean precipitation chance, temperature, weather change (baro change), wind velocity, and wind direction. Gust velocities are much harder to predict, but the presence of gusty conditions is predictable. Snow fall amounts, whether from lake effect or synoptic, is very difficult. I have seen heavy snow at one end of a shopping center and only a dusting at the other end.

As for football, well Cleveland was ahead at the half and won -- its first win of the year. What more proof do you need? Actually, I am mostly kidding about football. There are just too many unpredictable factors to render any prediction of score reliable. I saw an article once (60 Minutes?) in which the bookies in Las Vegas were interviewed. The conclusion was basically that they lose money or just break even on NFL betting. They do it because they are expected to make odds and other betting by the same gamblers (i.e., pull through business) makes up for the losses. I don't know one way or the other and have never bet on any game.

John

 

Did your team win today and beat the bookie odds,anyone.

 

I have no idea what the odds were, but I can't believe anyone thought the Cleveland Clowns would win.

It was actually a pretty good game -- not the usual game marred by turnovers.

John

 

You have to understand risk factors as well though, injuries, top athletes really put themselves out there, top performance means top risk, a single injury can flush all your existing data down the toilet if it's risk factors aren't taken into account. You can't look at a whole team in a game such as American football because single players can have dramatic effects if there's a total team synergy, it's impossible to statistically analyze this currently, but living systems that complex can't be fully modeled reliably though I'm sure you can get close enough to increase your odds, the question is can you increase your odds to the point where you can make any money.

American football fields 22 players. That is a lot when it comes to sports. Baseball teams field 9 and Basketball teams field 5. It seems to me the larger the number of players fielded, the less the team is affected by single injuries. But no matter what sport, if a team has enough injuries to its starters, it will probably degenerate.

Oh yeah, and betting on the NFL is illegal so if you ever did find such a system you'd end up in jail if you couldn't hide it.

I thought gambling was legal nowadays.

 

I think the basic idea behind casinos is that they put the pressure of winning on you such that if you don't win, they do. Take dice: You roll them once and get 6. Now what are the odds of rolling 6 again? And the casino doesn't even have to roll the dice. The odds are therefore greatly in the casino's favor. Even though you might win a particular round, probability says you'll lose the next one.

 

Bill, look up the book odds of being able to determine the Superbowl winner at the end of the year at the start of the season =)

Yes, they do know their buisness, and they'll never explain their statistical analysis methods =)

It would be hard to guess the Superbowl winner at the start of a season if all teams were equal. The odds would be 1 in 32. But past performance does count. I think you can at least boil the thing down to a half-dozen contenders. and if I'm right about that, it makes the odds 1 in 6. Looking over last year's data and the apparent trend in Green Bay's improved performance, I'm picking them.

 

Well, on weather, I think we will disagree. Cleveland is also near a Great Lake, and I find those aspects of weather that I expect to be predicted are predicted quite well. By that I mean precipitation chance, temperature, weather change (baro change), wind velocity, and wind direction. Gust velocities are much harder to predict, but the presence of gusty conditions is predictable. Snow fall amounts, whether from lake effect or synoptic, is very difficult. I have seen heavy snow at one end of a shopping center and only a dusting at the other end.

As for football, well Cleveland was ahead at the half and won -- its first win of the year. What more proof do you need? Actually, I am mostly kidding about football. There are just too many unpredictable factors to render any prediction of score reliable. I saw an article once (60 Minutes?) in which the bookies in Las Vegas were interviewed. The conclusion was basically that they lose money or just break even on NFL betting. They do it because they are expected to make odds and other betting by the same gamblers (i.e., pull through business) makes up for the losses. I don't know one way or the other and have never bet on any game.

John

I predicted the Giants would beat the Bears based on last year's performance. I think the Bears have overachieved in the first part of this year and the Giants have underachieved. The Giants were suffering from the hungry wolf syndrome while the Bears were smug about having beat Green Bay. It's hard to quantify the psychology factor, but it's certainly there. By the way I predicted the Cowboys would beat the Texans last week and for the same reason, except for the fact that the Texans are really good and not overachievers like the Bears. And concerning the Bengals-Browns game, last year's Bengals were overachievers. This is all based on my examining the statistics.

 

I have no idea what the odds were, but I can't believe anyone thought the Cleveland Clowns would win.

It was actually a pretty good game -- not the usual game marred by turnovers.

John

The Line was Bengals by 3. But the Pick was Browns 24, Bengals 20.

 

Here's an interesting statistic. In 2009 teams overall averaged 21.5 points on both offense and defense. They also averaged a gain of 325 yard on offense and a loss of 325 yards defense.

 

Isn't that expected after all? The teams play against each other, not against the "house".

 

That pointless statistic is only proof that a simple average of the whole is not representative of any factor which actually effects the game. There is a winner and a loser in the end, no average is ever a mirror in the NFL, check your numbers, and don't round. Someone won the super bowl that year, someone lost figure on the years average by how many points on a average they won by.

 

But thinking about the odds in terms of real math should shatter any delusions about syestems of gambling, unless there is more to it with respect to a certain game. The big time make-money gambling games are poker and blackjack. The rest of it is hopeless. There is a thing called card-counting that enables the odds in your favor. This is due to 7 people being present at a card game and the cards are all out there to count.

 

Pointless statistics seems to be the name of the game in the NFL nowadays. It is very hard nowadays to predict a winner. Perhaps by the end of the season this condition will change.

 

I can talk for poker to a certain extent. Odds are part of the science of the game, but good players can "bend" the odds in their favor, by taking advantage of other factors, like a player's gaming personality, intimidation and deceit.
Also, in poker tables in casinos the money at stake don't belong to the house and the house doesn't have a shot at them, unlike blackjack. However the house takes a part of the pot every round (in ring cash games) or at the beginning of the table (elimination games).

 

Isn't that expected after all? The teams play against each other, not against the "house".

Yes, it is expected. I think insurance companies look at the overall probabilities and win because of known overall odds of this or that disaster happening, though they may lose in individual cases.

 

There were many a great mathematician in the past, such as Pascal, who thought they could develop a system based on mathematics. It seems that some games lend themselves to this, but some don't. You would think that a game such as roulette would give decent --nearly 50% -- odds when betting on a color rather or on odd/even rather than a particular number, but the house still wins most of the time. I think this is a type of game that lures a player to place larger and larger bets until, after winning one round and loosing five, he's wiped out. If he was "wise" he put a stop loss number on his betting and walks away from it when he hits that mark, but odds are he had to try just one more time, and then one more, until he's maxing out his credit cards!