No one is saying they are set in stone, just that they are the best way to project/forecast. At the end of the day, win/loss is just a stat too. Just much less sophisticated in terms of forecasting value. If a particular projection is wildly off, that doesn't invalidate the entire field of statistics, just the model that was used in that particular instance. Numbers never lie. Numbers can't make mistakes, people do. That's the mistake DD is making, generalizing all of statistics just because a few people use it wrong.
there's a game scheduled for next week...make all the predictions you want...tell me about how the stats make it so unlikely for team A to win. i still don't care. I care whether Team A actually wins...and if they do, your stats didn't mean jack crap. All they did was help you formulate your best guess. Whooptie doo. i've seen it enough now to know that they're not as great at predicting outcome as many think they are.
i'm not blaming the numbers...i'm blaming the "worship at the altar" of statistics mindset that has us believing we can actually pinpoint how many games a team is going to win over 82 games (or 162 in baseball) based on stats from the past. we have a cottage industry around that concept right now. it's silly.
The title of this thread is "Basketball Prospectus Rockets Preview". If you don't care about forecasts, that's all well and good, but I don't know why you even bothered to click in and complain about something you don't believe in. Same goes for DD.
Because it is getting silly reading how much statistics are being considered these days. It is like folks think they can effectively tell you anything, when the reality is that they can't... They might be able to narrow down probabilities, but do you think the teams that left Rafer Alston wide open on the days he was hot cared when they lost the game because of that? DD
for fun!!!!!!!!!!!! i'm also a gigantic troll, which accounts for 49K+ posts. i care about forecasts...a little bit. i just don't buy into it.
The guy that watches the games and doesn't care for statistics at all will be just as accurate as the stat nerd when it comes to predicting. DD
The value of statistics or projections is not to predict the results next week's game. We are not bookies here. The result of that one game has more to do with luck than anything else. The value of statistics is to project over a broader sample size to gauge our strengths and weaknesses. So maybe we can start improving our team proactively, instead of waiting until our weaknesses smacks us in the face.
i'm good with that...actually i'm great with it! what i'm not as good with is the projection that "the Astros/Rockets will win X number of games this season because we put it in our formula and that's what it told us." that's a load of crap. and i see it all the time.
If I care about something *a little*, then I'll pay *a little* attention to those who really care about getting it right. I don't complain that those people care about it too much. People have different interests in life, after all. And your mistake here is that you assume the people at Basketball Prospectus believe their projections are pinpoint accurate. They don't. They're giving their forecast based on the numbers while they also recognize all the variables that aren't being taken into account. I think they understand the limitations of their models better than you do. If you actually bothered to read the multiple pages of text talking about the team instead of glancing at the numbers and rolling your eyes, maybe you'd understand their perspective better.
You said all that matters is wins and losses. So the proper question is this: will a guy who only looks at a team's win/loss record be as accurate as the "stat nerd"?
The goal should be to understand how and why the team played as they did. If you understand the manner in which they achieved their wins and losses, that will tell you much more about the team (and give you a better idea of how they'll perform in the future) than simply looking at the wins and losses itself. By your arguments, a team has no need to even track how well they are rebounding or defending the three-pointer or taking care of the ball, etc. Just count their wins and losses, which is all that matters, and that's the extent of the "analysis" needed. Come on -- I know you don't really believe that.
I like how someone who doesn't "believe" in stats so proudly wears a signature that displays his ignorance in statistical form. It's, like, fate.
Durv, I contend that observational data is just as accurate as advanced statistics in the ability to predict a games outcome. I can tell by watching the game what the problems are, rebouding, lack of passing etc....and what type of players the other team has and how they match up etc. So, how about a little season long challenge, you and I predict each game, we will both break them down in our own way and predict the outcomes. See who is actually more accurate.....I think it could be fun. It is making fun of those who overvalue statistics......thanks for playing. DD
