According to PowerStats, an organization devoted to soccer statistics, the Portland Timbers were better in almost every category during last Saturday's Cascadia Cup match against the Seattle Sounders.
Despite the 1-1 draw the Timbers made fewer mistakes, made fewer fouls, and won the ball more often than the Sounders.
about 1 year ago
Geoff Gibson
29 comments
1 recs |
Email
PrintComments
You might want to look into their metrics
Maybe the Timbers did play better, but most of their date is subjective.
Editor/writer at Sounder at Heart, MLS editor SB Nation. Follow me on Twitter
Just look into their "metrics"
They aren’t based on measurable things. They are based on whatever they want to base them on. I’m not saying they are irrelevant. But they are subjective measurements.
Editor/writer at Sounder at Heart, MLS editor SB Nation. Follow me on Twitter
by Jeremiah Oshan on May 18, 2011 3:37 PM PDT up reply actions
True to some extent
But isn’t giving an Assist in Basketball also subjective in nature? Or the awarding of a hit in Baseball by the official scorekeeper?
How about counting Shots in Soccer? Do you count the cross that the keeper had to save?
Contributing editor to Stumptown Footy the Portland Timbers SBN blog.
the awarding of a hit is very rarely subjective
I believe you’re thinking of the awarding of an error, which is entirely subjective A hit is only subjective when the scorer might have ruled it an error but didn’t. Anyway, Oshan has a point – and you reinforce it with the “counting shots” idea. I’m not sold on PowerStats just yet.
Your right
I meant error. A hit is a hit, that is easy.
I am not sold on them Either but taking only their data, which I hope is the same person doing the Portland Games, you can see a general trend upwards for the Portland Games which is what I find the most interesting.
Contributing editor to Stumptown Footy the Portland Timbers SBN blog.
Unless it's an error,
then it’s not a hit.
Expectation is the root of all heartache. — William Shakespeare
by TheDutchOven on May 19, 2011 8:08 PM PDT up reply actions
Hey now!
No creating circular arguments here, it will make my head explode.
Contributing editor to Stumptown Footy the Portland Timbers SBN blog.
Point taken. I did misunderstand your comment at first. I thought you were referring to the categories and not the measurements (which I didn’t even look at). Poor reading comprehension.
Keep in mind that there are subjective measurements that are much more reliable and relevant than some objective measurements. So just because something is based on “whatever they want to base them on”, doesn’t mean that it isn’t very useful (which may not be the case with PowerStats after all).
My whole problem with things like PowerStats is the same problem I have with official scorers in baseball
They’re different people at each park (I believe this is the case with PowerStats). They’re supposed to follow the same guidelines, but subjectivity is still a big part of it.
Or in Basketball
Where a players rebound numbers can be better at home vs on the road because one guy gives a rebound for a controlled tip while the others don’t.
Contributing editor to Stumptown Footy the Portland Timbers SBN blog.
Nothing is certain
but if soccer stats can start to be as reliable as baseball stats, PowerStats will have accomplished something huge.
I don't doubt that they're imperfect numbers
But I think these are better than either subjective impressions or the meagre few opta stats that get released.
Still, I’m a bit puzzled by some of their results- in the stats designed to measure effectiveness of possession, the Timbers won more 50-50 balls, produced more takeaways, suffered fewer giveaways, and committed fewer fouls, and still managed to get clobbered in Time of Possession. Maybe the stats are telling us something important- the Timbers spent their ToP trying to advance the ball and spent more time in positive action whereas the Sounders passed it around the back- but the result seems odd to me.
And while I find the stats they accumulate interesting in their own right, I’m skeptical of their aggregating mechanism. Why is scoring a goal worth the same number of points as three successful crosses (“reaches”)? Do they have stats to back up that on average 3 crosses will produce a goal? I think this can be interesting to compare styles or to eyeball a player’s contributions, but I’m skeptical that aggregating in this fashion reveals much. I’d love to see a fantasy league based around this system, and comparing individuals might be somewhat fruitful if you take the aggregation with a heaping dose of salt, but I don’t know about measuring team performance with it.
You neglect to mention that the match report claims an even match.
According to, http://www.powerstats101.com/article/141/facts_from_sounders_and_timbers_1_1_draw:
“Rivals Even on Scoreboard and on Stats Sheet…With anticipation running high, it’s no surprise that neither team was able to impose consistent pressure on the other. A look a the stats sheet shows us that 1-1 was a fitting scoreline for the Pacific Northwestern clubs.”
Some clarification
Hey guys, it’s my website you’re talking about.
First of all, let me start off by saying thank you for linking/talking about my site. It means a great deal, since it’s been a ton of work
Second of all, let me clarify a few things.
1) the ratings are not subjective at all but based on specific actions on the pitch. We start with the core 20 stats and then things like Attacking Rating, Striker Rating, Defender Rating, etc are based on a simple formula as described in a our FAQ,.
We don’t apply an arbitrary PowerRating to who we ‘felt’ did well, but it is a tally of their point accumulation throughout the game.
It gets no more objective.
2) As for the ‘why is a goal only worth 3 points’ question, this is probably the most common question/concern we get. This system has been worked on for quite a long time and that was initially one of the things we worked on the most. What we found is that if you weigh the goal too much, then the rest of the stats are pretty worthless.
What we’re trying to do is show you who the best player on the pitch was. If you believe the guy who scores the winning goal is the best player regardless of anything else, then stats aren’t for you. Likewise for crucial save of the match.
With our system you don’t have to have a goal, assist, save to be the best player on the pitch. And we think that is what is meaningful.
3) I did all of the Portland games so the consistency is there.
4) We’ll be doing all of the Portland games from here on out so be sure to check back each week and see if you agree with the ratings.
Not sure how what you're saying doesn't point to subjectivity
What 20 stats do you use, for instance. How do you determine what a positive pass is? How do you determine what a negative pass is? Just because you have a point system doesn’t make it objective. If you’ve really managed to make an objective scoring system, that should be commended, but without knowing all the ins and outs, this doesn’t seem anymore objective than the Castrol rankings.
Editor/writer at Sounder at Heart, MLS editor SB Nation. Follow me on Twitter
by Jeremiah Oshan on May 19, 2011 11:11 AM PDT up reply actions
Well we don’t actually deal with touches, passes, possession as you think in the traditional sense. Those figures can easily be manipulated by useless defensive possession and back passing. What we do is only log stats based on the two criteria
1) What happens at the change of possession
2) What happens around the box.
Really, aside from scoring, those are the two most important aspects of the game.
So our stats really are objective. For example, a player losses possession without being forced, giveaway. If a defender knocks the ball away and his team gains possession, take-away.
There’s no guess work on our end where we say ‘well his idea to cross into space was good, even though now on got there.’ That’s fine, but it’s still a giveaway.
As for the Castrol ratings, our system is completely transparent and that’s the way we want it.
Thanks for chiming in!
A couple clarifications, if I might ask:
the ratings are not subjective at all but based on specific actions on the pitch
I think that’s understood, but my question is more: are there standardized definitions of what those actions are? To take an outlier example, if a mishit cross goes in the net untouched by another player, does it count as a cross or as a shot? How do you define the actions?
I did all of the Portland games so the consistency is there
But if other people do the other games, and interpret actions differently than yourself, you can’t compare across teams. Have you taken steps to ensure this won’t happen?
PowerStats is a good start and I hope it develops more, and I understand if some of your methodology is proprietary/unpublishable, but it would be interesting to hear some more background if you’re willing to share.
An example or two, from your FAQ
you define “shot” as:
An intentional attempt to score a GOAL initiated by the offensive team, and where the ball comes relatively close to the goal.
What is “relatively close”? Does each PowerStats scorer interpret that differently, or is there a number associated with “relatively close”?
You define “reach” as:
Physically directing the ball into the penalty area where there is a scoring opportunity
Does that include a defensive player who turns the ball over to an opposing player who then shoots?
yes there are very definable rules for what constitutes a shot and reach. We’ve created a training manual that any stat guy has to adhere to the official definitions of a stats.
Relatively close has a list of definitions depending on where on the pitch the shot was taken.
Generally, if the shot goes 6 yards over the crossbar it’s not a shot. If it goes wide of the six yard box, also not a shot
Would you be willing to share your training manual so we can see your definitions?
I’m really not trying to be a jerk, I just would love to understand the definitions you’re using. If your training manual is proprietary, no worries, but I thought I’d ask.
Don't get me wrong
I think this is a good start in the quest for better soccer stats, I just think we need to be as opaque as possible when talking about them.
Editor/writer at Sounder at Heart, MLS editor SB Nation. Follow me on Twitter
by Jeremiah Oshan on May 19, 2011 11:16 AM PDT up reply actions
you mean transparent
and it is our direct goal to be as transparent as possible
Thanks!
Thank you for explaining a lot of methodology you use and that you do have a training manual. I love the stats and love to see a different look into Stats and Soccer.
My Brother and I discuss the your stats almost weekly so keep it up!
Contributing editor to Stumptown Footy the Portland Timbers SBN blog.
After looking inside a little deeper
I must say that I like where this is going. I’m still not sure I agree with the exact scoring system (a goal is 3 points, an assist 2, for instance) but it does get credit for doing something no one else is doing and this is quantifying soccer in a new way.
Editor/writer at Sounder at Heart, MLS editor SB Nation. Follow me on Twitter
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
