Champion data v Pro-stats

[joe444]

In my limited experience, these are the 'big two' sources of stats. Pro-stats seem easier to access; Chapiondata seem to be a bit secretive but more trusted. I'm not sure who provides the AFL site with their numbers.

So, which is better? Why? If this is an obvious newbie question, sorry.

 

[joe444]

But for the meantime, as this valid discussion develops. Here is how Champion Data regard a CONTESTED POSSESSION:

a kick by a player to an opposition player on the bounce is considered a loose ball get and therefore recorded as a contested possession;

a kick by a player to an opposition player is marked by an opponent, it's considered an uncontested mark and therefore an uncontested possession

Pro-stats, as I understand it, review their stats after the game by multiple viewings of the tape. CData provide live stats to the AFL. Doesn't make them better.

 

[Belnakor]

In my limited experience, these are the 'big two' sources of stats. Pro-stats seem easier to access; Chapiondata seem to be a bit secretive but more trusted. I'm not sure who provides the AFL site with their numbers.

Champion data are the "official" stats of the AFL. However the public does not have access because they want to flog the stats to the AFL with a much bigger markup.

Obviously its too complicated for them to provide the stats to joe public, and have a special premium offer (such as real time stats) for the teams.

IMHO they are no better than pro-stats, and quite possibly could be considerably worse. We have no idea of the methology Champion Data use, and your example is good example of how their stats are quite likely flawed. They have no external verification of their stats (or if they do, the public are unaware of this).

pro-stats is good for joe public, and imho their stats are as reliable as Champion Data (though in reality you are taking both pro-stats and champion on faith, perhaps Champion more so).

Champion do seem to have some more interesting stats hidden from the public, such as meters gained which we are unfortunately not allowed to access. Taps to advantage is another.

 

[joe444]

We have no idea of the methology Champion Data use, and your example is good example of how their stats are quite likely flawed. They have no external verification of their stats (or if they do, the public are unaware of this).

This strikes me as an important point. CD have the advantage of a sweet contract, special articles in the HUN (whom founder Ted Hopkins used to write for), but they seem very secretive about how the methodology somehow ensures their stats collated live are superior to those garnered from multiple viewings on tape.

 

[Champion_Data]

Hi folks. I'm not going to get into the "who is better" debate, but for the record we have 10 people working live on game day to ensure the accuracy of our stats.

All of the AFL clubs use our data for vision editing so if there has been blatant errors in the past I doubt we'd still have a job today.

We are anything but secretive, but due to our limited contact with the general public I understand it could appear that way. This year's edition of the AFL Prospectus (shameless self-promotion) has a snippet about our game day structure and there's a video on Youtube of an Access All Areas segment we did with One Week At A Time a few years back here.

p.s. I'd personally take Bartel over Lewis...

 

[Belnakor]

the AFL clubs would have at the most 3-4 people working on statistics and probably noone working in the analytics space, so forgive me if I think that the "AFL clubs think its ok" is taken with a grain of salt.

 

[Belnakor]

We are anything but secretive, but due to our limited contact with the general public I understand it could appear that way.

This bit made me laugh. If the stats are so good why not give the general public access to the database for a fixed fee?

btw i'm commenting as if this guy is legit when he is highly likely not.

 

[Rourke]

the AFL clubs would have at the most 3-4 people working on statistics and probably noone working in the analytics space, so forgive me if I think that the "AFL clubs think its ok" is taken with a grain of salt.

You're about right with your assumption, but one thing the clubs do care about is accurately identifying players. I used to work for Champion but am now in competition, so you can take my view with some grains of salt too. One club's feedback was that even their club stats-gatherers couldn't tell the difference between [three defenders] as accurately as the Champion callers, and when they found a discrepancy Champion was nearly always right.

There are benefits and pitfalls from using video only. Especially Channel 7 footage misses many touches altogether, and Champion has the advantage of an at-ground spotter. However doing it live puts a strain on the accuracy off-the-ball stats like shepherds.

This bit made me laugh. If the stats are so good why not give the general public access to the database for a fixed fee?

btw i'm commenting as if this guy is legit when he is highly likely not.

Even though it's a single-post account, I reckon it's legit. Is that you Doctor?

They do offer stats for a fee, albeit an exorbitant fee unless it's for pure research from a Uni etc. Not for me to talk about their commercial approach ...

 

[doppleganger]

Champion has the advantage of an at-ground spotter. However doing it live puts a strain on the accuracy off-the-ball stats like shepherds.

The at the ground spotter can help with your basic kicks, marks handballs or perhaps determining whether a mark was given or a free kick was awarded (even the commentators get this wrong sometimes).

However the more speculative stats like tackles, are much easier viewed with the aid of a replay. The inconsistency with CDs tackle counts baffles, especially when you are watching with SC or DT in mind, you often see an opponents player get awarded a tackle for simply being part of a pack. It seems there is inconsistency between different states in particularly with tackles.

Getting into bed with the AFL commercially has all but removed any comparison anyway, as they want to promote the one 'official' stats.

 

[Rourke]

The at the ground spotter can help with your basic kicks, marks handballs or perhaps determining whether a mark was given or a free kick was awarded (even the commentators get this wrong sometimes).

However the more speculative stats like tackles, are much easier viewed with the aid of a replay. The inconsistency with CDs tackle counts baffles, especially when you are watching with SC or DT in mind, you often see an opponents player get awarded a tackle for simply being part of a pack. It seems there is inconsistency between different states in particularly with tackles.

Agree with you on tackles, but the trouble is every coach/club defines tackles differently as well. Champion has a consistent 'middle ground' definition that looks at whether the tackle materially affected the quality of the ball-carrier's disposal, but even that can lead to some bizarre-looking calls. This is an area clubs scrutinise heavily themselves, because they're interested in players' defensive workrates and the official tackle stats don't reflect that in any way.

 

[doppleganger]

Agree with you on tackles, but the trouble is every coach/club defines tackles differently as well. Champion has a consistent 'middle ground' definition that looks at whether the tackle materially affected the quality of the ball-carrier's disposal, but even that can lead to some bizarre-looking calls. This is an area clubs scrutinise heavily themselves, because they're interested in players' defensive workrates and the official tackle stats don't reflect that in any way.

Yes am aware of that, but it is frustrating that fantasy comps can often be decided by which state and Champion callers are watching a game!

This is partly why the AFL has pressured media outlets to use the 'official' Champion Stats, just to have the one version out there.

Even with disposals, often a few minor disagreements between Pro and CD, usually around disputed kicks off the ground, or questionable handballs....who was correct depends on intepretation etc, but the slight inconsistency didn't appeal to the AFL.

 

[Gibbsy]

This bit made me laugh. If the stats are so good why not give the general public access to the database for a fixed fee?

btw i'm commenting as if this guy is legit when he is highly likely not.

Imo he is legit because there's no team under his name, only special people can request that now I think

 

[Footynomics]

Can't say I've ever really understood why the AFL don't make as many stats as possible public. Surely having as much information out there as possible can only be good for interest in the game. Look at Major League Baseball, it took a while but they finally got the got the stats out there and it has revolutionized how some teams conduct their operations. All started by the fans who took the time to understand the stats and how the game should really be played. Imagine footy did the same, there's plenty of people with enough interest in how the game works and how it could be made better to do something useful with additional information. As it is we have to make do with snippets as supplied on a drip feed by Champion Data. While their statisticians might have some great ideas on how to improve player evaluation and strategy I'm sure they haven't thought of everything.

 

[pmalcolm1984]

Hi folks. I'm not going to get into the "who is better" debate, but for the record we have 10 people working live on game day to ensure the accuracy of our stats.

All of the AFL clubs use our data for vision editing so if there has been blatant errors in the past I doubt we'd still have a job today.

We are anything but secretive, but due to our limited contact with the general public I understand it could appear that way. This year's edition of the AFL Prospectus (shameless self-promotion) has a snippet about our game day structure and there's a video on Youtube of an Access All Areas segment we did with One Week At A Time a few years back here.

p.s. I'd personally take Bartel over Lewis...

Champion_Data, given that you have effectively driven away any form of competition from the market, I'm guessing you'll be developing a subscription based website like Prowess did in pro-stats.com.au?

 

[Smiling Buddha]

Are there legalities surrounding the 'creation'/use of statistics from AFL matches?

Although I wouldn't think there should be such a law, it wouldn't surprise me if there were some law which allowed the AFL to claim 'ownership' of statistics arising from AFL matches, no matter who compiles those stats or how.

That is, if Joe Bloggs started a website called Bloggstats and offered a subscription service whereby customers paid to access statistics Joe compiled from watching the games and recording the stats himself, can/does the AFL try to shut Joe down (in the interest of protecting their commercial partner i.e. ChampionData)?

Is this why pro-stats was down for a while?

 

[Rourke]

Are there legalities surrounding the 'creation'/use of statistics from AFL matches?

Although I wouldn't think there should be such a law, it wouldn't surprise me if there were some law which allowed the AFL to claim 'ownership' of statistics arising from AFL matches, no matter who compiles those stats or how.

That is, if Joe Bloggs started a website called Bloggstats and offered a subscription service whereby customers paid to access statistics Joe compiled from watching the games and recording the stats himself, can/does the AFL try to shut Joe down (in the interest of protecting their commercial partner i.e. ChampionData)?

The legalities have never been tested. The AFL might claim ownership (as MLB does for all products derived from the telecast of baseball in the US) but that claim could well fail. Champion Data pays a licence fee to the AFL, and is partly owned by it, which means it gets some level of protection. The AFL could also chip away at the edges, protecting their fixture, team emblems, attempting to enforce ticket conditions etc to make it harder to compete. In summary, pay for some real legal advice if you're serious

Is this why pro-stats was down for a while?

Not exactly. My understanding is that this was not due to a claim of ownership.

 

[TMoney22]

Do clubs, media, AFL have software to browse stats gathered by champion data, or is it a request service where the required stats are provided in excel?

 

[Rourke]

Do clubs, media, AFL have software to browse stats gathered by champion data, or is it a request service where the required stats are provided in excel?

They are provided with custom software to browse live stats and season reports. You can sometimes see it on a screen behind the commentators during the breaks.

 

[TMoney22]

They are provided with custom software to browse live stats and season reports. You can sometimes see it on a screen behind the commentators during the breaks.

OK cool. I do recall seeing that, but very specific and graphical which is nice. But I'd imagine the data/stats would be held in table form and difficult to pull big data quickly.
for example. A teams disposal efficiency over the last 3 years by player by stadium by quarter. or something like that

 

[Belnakor]

Do clubs, media, AFL have software to browse stats gathered by champion data, or is it a request service where the required stats are provided in excel?

doubt they have backend access, would be a front end query tool at best.

 

[Zimpy27]

Does anyone know if they offer stats on ground position to the public?

Also stats on individual matchups would be very handy?


When they record games do they use stop-motion cameras? Simple images captured in timed intervals can be processed by algorithms quite effectively. If you had a camera at the ground you could map the ground from an angle and use it to create a grid to view the locales of players at all times. Surely something like this exists?

Or do players all where trackers now?