bsilver wrote:payitforward wrote:bsilver wrote:Projecting is only useful if there's sufficient data. A projection made after the 6-12 start would have been for a 28-44 season. Who's to say the 20-19 current record is adequate for a projection?
Also ignored is team improvement. After the 6-12 start, they are 14-7. Maybe the 14-7 Wizards are the "real" Wizards that reflect team improvements made by coaches/players. There was bound to be an adjustment period with a new coach and many new players. If we ignore the 6-12 start and say the 14-7 team is the true team, then the record projects out to be about 46-36.
Also ignored is any future improvements. Maybe Mahinmi comes back and has an impact. Maybe a signing or trade improves our horrible bench.
Projections are useless unless previous data is sufficient, and there will not be future changes.
Not to get political, but what if we looked at global temperatures over the last hundred years. A projection would show temperatures rising at a much more modest rate than if we looked at the last 20 years. Maybe the best projection is to look at recent temperatures, and maybe the best Wizard projection is to look at more recent performance.
Sigh...
What you mean is that projections aren't predictive. They aren't "useful" in that sense. Which is what I said in the post you'd like to argue with.
You are right: "Maybe Mahinmi comes back and has an impact" or maybe he comes back & has no impact. Or maybe he doesn't come back. Or maybe we make a trade & get better. Or worse. Or suffer an injury. Or...
But, that doesn't mean projections are useless. Let me demonstrate: "After the 6-12 start", we were 5-3 at home, which projects to 25-26 home wins, & 1-9 on the road, which projects to 4-5 road wins. IOW, we projected to a max 31 wins.
Suppose we include 1 more game -- we beat the Nets on their court. At 7-12, we are still 5-3 at home (25-6 wins), but now we are 2-9 on the road, which projects to 7-8 wins -- we project to a 34 win season.
In other words, our 13-7 record since the Nets win has lifted us to a projection of 4 more wins. Not that big a difference, huh? & why is that? Because 12 of those 20 games were at home.
Projections aren't predictions. Projections are information. Useful information.
Sigh, sigh, sigh...
There's an actually an eye-roll smilie so I'll add that to the argument
What if the starting point of the projection is the beginning of the 11 game home win streak? I imagine that projection would turn out pretty good. 11-0 at home, and 2-6 on the road. We'll win all 20 remaining home games, and go 6-17 (or 5-18) on the road for the 23 road games, for a final record of 45 or 46 wins.
So, we have competing projections - one based on the entire season's worth of data, and another using the last 19 games.
IMO projections are only useful information when combined with analysis. What if Lebron missed the first 20 games of season and the Cavs went 5-15 and when he returned they went 15-5, for a 20-20 record. The projection would be a 41-41 season. Not very useful. An analysis of the data would lead to throwing out the first 20 games to make a worthwhile projection.
As others have mentioned, it's common for NBA teams to improve, and the Wiz certainly have. In these cases the early part of the season should be thrown out when trying to make a useful projection.
Once again, you repeat my point as if repeating it disproved it. Yes, you are right, cherry picking the start & end of the data set will give you different projections. That was what I said!
Then you point out that projections are not predictive (your LeBron story). That too was what I said!
Then you said something completely idiotic: "it's common for NBA teams to improve". Think about it. NBA teams "improve" by increasing their winning %. But, NBA teams play other NBA teams. It can't be more common for teams to improve than it is for them to get worse.
Finally (& I do mean finally; if you continue on this subject I'm putting you on ignore, which I don't want to do b/c you write perfectly sensibly most of the time), the point of my projection was only to highlight the fact that our 20-19 record should be seen in the light of our having played way more games at home than on the road. That's all. The projection simply illustrates what happens when you do see it in that light. It makes it easy to see the effect of that fact. That's what is "useful" about it. That & only that.
In other words, the function of the projection is, to use your word, "analysis" -- not prediction. Clear? Now, please drop it.