number talks

I was talking about basketball with some people online these few days. One of the stats of one player became a topic, Assist.

The point I was trying to argue was that assist of a player meant more than his willingness to share the ball. It also meant how often a player possess the ball on court. For example, a point guard are usually the ones who brought the ball up the court. So it is natural for that position to have more assist than most other players. It also applied to the star players, because they also have more chance to possess the ball. The more chance a player could put his fingers on ball, the more freedom for him to do as he wish. A star player in this case can put the ball on the ground, driving to basket; or he could shoot on the spot; or he could pass to make an assist.

The more I thought about this, the more I dealt deeper into the stats, for games, for players. There are many interesting co-relations within a player's stats. Like the assist(apg) and scoring(ppg). It is indeed true that most of high scoring players do have more assists than most other players.

But it is more interesting about the game stats. What I meant by game stats is that, after each game, there are more than three accounts of that game on the NBA recap page. The most obvious are, game story, box score, and play-by-play. If I just looked at box score, it seemed it provided a complete picture of how the team and each player are performed in one game. But it does not necessarily reflect the game results. It seemed that in a close game, a play-by-play may be more accurately reflect the game results.

If I looked at game stats, I can see how each player and each team performed, across the whole game. But it does not tell how do one react in the fourth quarter? A player could perform badly the whole night, except in the final minutes. The chance of that may be surprising, as the game draw near, and each player's focus is up. So these two stats tells something different about games. In a purely stats sense, box score is a pure stats. Play-by-play is more like a chronicle, which provided with much less stats, but more narration. By tracking box score of each game of each team, I might be able to learn how a team is performing. For example, I could look at assist stats of each team and learn about the efficiency of their play. Because I believe that apg of a team tells me how aggressive the team is running the basket. By studying the play-by-play, I learned each player's flow, do they heat up fast; or do they start slow? Who is more focused, who performed worse...

This leads me to wonder about how we could approach the history study. How should we view about stats from history. And what kind of stats should we collect? How we could use the timeline as a tool, which we need to ask, what should we use it for? The stats at one point of time, how doe it fit in a timeline? What kind of relationship can we draw between each stats?