Statistics vs Watching the Games

[Jimmy76]

What are the values of each and how do those values weigh against each other? What do statistics pick up that watching games doesn't? What does watching the games pick up that statistics don't? Does the objectivity of statistics outweigh the merits of watching a game that is considered too organic to asses perfectly through numbers?

some questions to spur the general discussion

and inb4 "need both" that a fine proposition but what is each one providing that the other isn't?

[Jakay]

Statistics are a product of the game and not vice versa.

My thoughts on the subject.

[Official]

Statistics, when analyzed with other statistics, is the best measure. One statistic by itself may have limited meaning.

However statistics comparing eras in basketball, even using advanced statistics, is better left alone.

[wreck]

Statistics are a product of the game and not vice versa.

My thoughts on the subject.

Correct. Statistics are a numerical summarization of what happened in the game and help disprove (or support) what most people saw during the game.

[wreck]

What are the values of each and how do those values weigh against each other? What do statistics pick up that watching games doesn't? What does watching the games pick up that statistics don't? Does the objectivity of statistics outweigh the merits of watching a game that is considered too organic to asses perfectly through numbers?

some questions to spur the general discussion

and inb4 "need both" that a fine proposition but what is each one providing that the other isn't?

I can tell that this thread will turn out to be a debate thread over which players are better because that statistics show one thing whereas a subjective eye shows another thing.

Statistics are definitely a more accurate judge of production than most humans so I don't see why they are so discredited on this board. Then again, I'm pretty sure I can think of one example

[Doctor MJ]

Depends on who is doing the watching, who is using the stats, and how each is being applied.

The better you are at watching the game, the more important it is you watch.

The better you are with math/analytical stuff, the more important it is you use stats.

Last, the bigger the timescale, the harder it is to watch all the games, and the worse the human brain is at balancing positives and negatives, so it becomes more important that you use stats. What that amounts to is that if you're watching a series and evaluating that, you can do a lot without stats. If you're watching one team all year, you can do a lot evaluating that team without stats - and conversely it's pretty silly to think you understand what happened all that well compared to a knowledgeable fan who did watch. Anything beyond that, if you aren't making stats a big part of your analysis, you probably don't understand things as well as you think you do.

[KevinMcreynolds]

You need a combination of both. You would look at Josh Smith's stats and think he's a great player, but then you watch a game and he's mailing it in in the 2nd quarter, rofl.

[Perfundle]

Statistics are a product of the game and not vice versa.

My thoughts on the subject.

Correct. Statistics are a numerical summarization of what happened in the game and help disprove (or support) what most people saw during the game.

Not true. Coaches use statistics all the time to change the gameplan for the next game. Certain trends can shape offensive and defensive strategies, such as the success rate of different offensive sets, the spots on the floor that opponents like to take shots at, the shooters that should go uncovered in favor of doubling others, etc. These are things that the ordinary human mind cannot process fast enough to remember.

Of course, statistics are definitely a product of the game, since you can't record stats without watching the game. And since the human mind can't recall stats particularly quickly, players are not calculating the probability that they switch to strategy B if the opponent deploys strategy C against strategy A; that sort of thing is left to the coaches.

Another problem with stats is how misleading they can be. Some stats weigh offense too heavily, some don't take into account the personnel on the floor, some unfairly punish fast-pace teams, and so on. And if everything is taken into account, the sample size might become too small. Finally, there's random variability that can skew stats without actually meaning anything significant. Statistics are most useful if you understand all their limitations.

[mystir]

Statistics:
- unbias
- available for every player in every game

Watching:
- context (difficulty of a rebound; how much attention a player is getting)
- better understanding (team A matches up well against team B, player X struggles against long defenders)

[mystir]

I don't like the "statistics are misleading" argument because its really no different when watching games.

Example:

Game 5 of the Cleveland/Boston series. Kenny Smith was talking about Rondo like he was playing well when he was 0-2 or 0-3 at the half with more turnovers than assists and generally uninvolved. He said it was because Rondo was setting the tempo. To Kenny Smith, I suppose that mattered, but its not a very objective thing where you can prove or disprove the value in it. So watching the games can be very misleading too, as far as using that as a basis of comparison.

Also, lets not forget the how much personal bias can factor into the equation.

[chubby_1_kenobi]

Both are needed to evaluate a player objectively. Relying too much on one or the other is faulty because a perfect statistical model of a basketball game does not exist and quite possibly never will exist.

When someone is able to develop a statistical method to differentiate the value of a basket made at the beginning of the game and a basket made to win the game, then it is time to stop watching the game and start watching the boxscore instead.

I mean, most statistical model count a rebound and an assist under the same value for christ sake. That's why most statistical model out there is biased toward players who rebound a lot (ie: big men). I'm not discounting the value of a rebound from a big man but if you tell me that 1 rebound holds the same value as 1 assist, I'm drop kicking your sorry arse.

[sp6r=underrated]

Depends on who is doing the watching, who is using the stats, and how each is being applied.

The better you are at watching the game, the more important it is you watch.

The better you are with math/analytical stuff, the more important it is you use stats.

Last, the bigger the timescale, the harder it is to watch all the games, and the worse the human brain is at balancing positives and negatives, so it becomes more important that you use stats. What that amounts to is that if you're watching a series and evaluating that, you can do a lot without stats. If you're watching one team all year, you can do a lot evaluating that team without stats - and conversely it's pretty silly to think you understand what happened all that well compared to a knowledgeable fan who did watch. Anything beyond that, if you aren't making stats a big part of your analysis, you probably don't understand things as well as you think you do.

+ 1

[houston911]

NOTHING is more annoying than a box score whore

A box score whore is one of you guys who always wants to talk about basketball, but for some reason you never watch games, you just stat track

Watching games>>>>>>>>>>>>any of those nerd ass stats that yall love so much

If you disagree then you're stupid. I can point out misleading stats from every game. Stats don't tell you about how someone plays on ball defense or rotates. Stats don't let you know who breaks down a defense or draws a double team. Stats don't show when someone gambles for a steal and misses and gives up a basket. FG% is misleading. Player A might have a higher fg percentage than player B, but stats don't show you that teams dare player A to shoot and run double teams at player b


basically yall need to order nba league pass and stop with this stat bullsh*t

the most knowledgeable basketball people I know don't discuss stats too much. People that know the game know that that stuff is misleading.

[GameOver25]

Observation and knowledge on how the game is played is key, stats should be to backup your argument or make a prediction.

[supaflash]

Stats are black and white, the game is not. That and we don't have stats that measure all sorts of things that happen in a game. And even some of the ones we have are somewhat arbitrary (assists anyone?) Basketball is such a flowing dynamic game that cut and dry stats inherently fall short of telling the story. If you like stats, go watch baseball, it is a much more cut and dry, stat based game, and even there they don't always tell the story.

Sports in general are not played for stats, competition is based on human effort, mentally and physically. When you can start quantifying people based on statistics, then you can start accepting them at face value for sports.

They are great for analysis, review, etc, but for now we still need a keen knowledgeable eye to discern truth behind the stats and the game, and even then there are factors that sometimes can't been seen or are unknown, game plans for example.

Lastly, if we could come up with truly fully accurate stats and comparisons, like player X made play Y against defender rated Z, plaing M rated defense on said play, with these players aorund, on this game plan, running this set vs this defense, at this point in game with these fatigue levels, etc etc etc... I think both the observers and the stat geeks alike would see a whole new world of how players, teams, and accomplishments stack up...

[Hendrix]

Depends on who is doing the watching, who is using the stats, and how each is being.

agreed

I'll take stats over a lot of people's opinions from watching the game. And by stats I don't mean raw box scores.

Raptor fans for example seem to have been fooled by J.O, and Turk from watching the odd game even though stats said otherwise. Then when they got to see them every game the stats started to make sense.

Also I've noticed a lot of other teams fans form negative opinions about Bosh if he had a bad/mediocre/injured game against them. They just don't see enough of him to form a good opinion, so would be better served with stats. And a PER of 25 for the season, and 28 pre all star break/injury is nothing to sneeze at.

[Dr Positivity]

I'll actually go with Option C: Skillsets

Every play brings a particular brand of skillset to the game. This is backed up by watching or stats equally. But ideally - we shouldn't have to rely on either to know whether an offense is set up to be succesful or not, if our understanding of the game is enough.

Evaluating defense is more a matter of watching and stats than skillset, though

[Harmless]

Humans are notoriously bad at remaining unbiased, and tend to color their opinions and conclusions (especially when it comes to players they love or hate). Statistical analysis, if done properly, is better at telling you how good a player actually is.

Remember, I added the qualifier "if done properly" because this is very important when looking at stats. Forget the "lies, damn lies and statistics" BS. Stats don't lie. It's the ones who manipulate them who do.

However, this does not mean you can just look at stats and ignore watching the games. You need to do both. There are many things that do not show up on a basketball stat sheet, especially on the defensive end. Who records tipped passes? Who records how many times you have a hand in the shooters face? Who records rebounds tipped to teammates? Is the player injured or sick at the time? None of those show up in the stat sheet.

[Finn Man]

I too think that you can't go with either one only.
You need to have your own opinion on player's talent/effort level by watching him or her play for a good amount of games, not just a sample size like two or three games.
When you have the knowledge on player's overall talent, then you must have something to prove that player A is better than player B, and that is usually stats, which (when read properly) can be used as good evidence.

I think in future, with the SYNERGY thing (a great invention!) you will be able to analyze players better, and not just going with stats also on these forums.

[Ortho Stice]

I don't like the "statistics are misleading" argument because its really no different when watching games.

Pretty much everyone knew last season that Durant was a great player who would only improve, but a statistician on the Mavericks had some formula which said he was one of the worst players in the league, and that if his team were offered Durant for free he wouldn't take him. This just shows how some statisticians can be worse basketball analysts than a normal five year old basketball fan who watches games.