Video Capture Just For Goalkeepers

By on February 25, 2021 in Videos

We recently had the pleasure of running a product demonstration of our new “computer vision” beta software at Volta DemoCamp 2021 – working name OptaVid. What is OptaVid? It’s an online algorithm that leverages your expertise as a coach or player, using it to automatically scan game footage to identify the parts of the video that show the goal area. It’s still in early development but we believe it has the potential to be a powerful tool for saving coaches time. In tests, we’ve found that it can condense a typical 90 minute game down to an average of 20 minutes or so of goalkeeper footage.

If you’re curious how it works, check out the clip above!

Ramping up our game

By on February 4, 2020 in Software Updates

It’s been a busy few months in the goalkeeping world. Not just in the European leagues and MSMSL pre-season – here at Stopper, we’ve been adding new features and functionality to the app, with plenty more in the pipeline to come!

Here’s a few of the update highlights from the latest iOS/Android Stopper release:

• In-progress games auto-save to prevent data loss in case of a crash
• NEW Communication tracking interface and stats
• NEW Goalkeeper self-evaluation interface and stats
• NEW Season stats updated to show Win-Tie-Loss record

Some are by popular request, while others have been suggestions from our awesome community of partners and users, all with the goal of making Stopper an even better training and development tool for high-performance keepers. Some of the most exciting changes coming are ones that are largely invisible – database configuration and back-end functionality that allows Stopper to connect to the beta of the Stopper Pro platform. Ultimately, this will allow coaches to easily track multiple players and analyze their game performance across a range of detailed metrics.

Pairing Video With Stopper Data

By on April 22, 2019 in Uncategorized

Integrating video feedback into a high-performance training environment helps both players and coaches improve. A 2010 study found that athletes supported with video feedback reported less anxiety and perceived their coaches far more positively than those who didn’t – meaning that video feedback not only helps improve technique, it also helps increase confidence and mental strength. It’s literally a game changer in terms of high-performance player development.

Over the last year, we’ve been working on a couple of innovative ways to incorporate video into our Stopper Pro beta.

First, we’ve been using the timecodes generated by tracking games with Stopper to auto-generate short 30 second video clips from the game footage. This has the potential to save coaches who use video feedback a lot of time – instead of scrubbing through a full game, they can review just the relevant sections of game footage. Even better, that footage uses the Stopper metadata to neatly organize each clip by type of goalkeeper action. It’s a great way to contextualize Stopper player data and also makes a great library for viewing, say, a season’s worth of 1v1 saves.

Secondly, we want to use video to make data collection easier in the first place. The beta version of our software identifies when the ball is in the defensive third, meaning that instead of collecting Stopper data live during a game (a challenge if you’re coaching multiple players) coaches can review the game footage and then simply jump to the parts of the game where the goalkeeper is likely to be part of the play. Over time we hope to introduce machine learning to the algorithm, meaning that not only will it identify when the ball is in the defensive third, but also when the goalkeeper makes a save or plays the ball. Ultimately, our hope is to bring the technology and data collection that’s available to professional teams within reach of high-performance development leagues in a cost effective, user friendly software package.

Finally, that study on video feedback found it wasn’t just great for players – coaches that used video feedback viewed their own performance more positively than coaches who didn’t. Seems like a win for everyone on the pitch!

The Data You Don’t See

By on March 6, 2019 in Uncategorized

This may be one of the coolest data analysis stories I’ve come across yet:

During WWII, the US Navy tried to determine where they needed to armor their aircraft to ensure they came back home. They ran an analysis of where planes had been shot up, and came up with a detailed map of bullet impacts. As the image below shows, obviously the places that needed to be up-armored are the ones taking most of the bullets: wingtips, the central body, and the elevators. That’s where most of the damage was happened.

Abraham Wald, a statistician, disagreed. He thought they should add better armor to the nose area, engines, and mid-body. Which was crazy, of course. That’s not where the planes were getting shot. Except Mr. Wald realized what the others didn’t – the planes were getting shot there too, but they weren’t making it home. What the Navy thought it had done was analyze where aircraft were suffering the most damage. What they had actually done was analyze where aircraft could suffer the most damage without catastrophic failure. All of the places that weren’t hit? Those planes had been shot there and crashed. They weren’t looking at the whole sample set, only the survivors.

Moral of the story: the numbers don’t lie, but our interpretation of them often hides the real truth.

Home Grown Talent

By on January 13, 2019 in Uncategorized

In 2014, the Canadian Hockey League banned European goalkeepers from playing in any of the CHL’s junior development leagues. Why? The intent was to kickstart homegrown Canadian goalkeeper development – by eliminating outside competition and hopefully, forcing teams to invest in developing local talent.

The ever-insightful Bill Reno of everybodysoccer.com (see link:.https://www.everybodysoccer.com/even-the-goalkeepers-like-to/2018/8/15/can-the-uswnt-learn-anything-from-the-canadian-hockey-league ) analyzes whether a similar approach might help the USWNT develop a new generation of top-tier goalkeepers. His conclusion is that the USSF ultimately needs to implement a complete system – from top-down leadership, to programming and training sessions at the grassroots level – if they wish to consistently develop positive results. To his point, simply removing tougher competition or relying on individual coaches to figure it out on their own is hardly a recipe for success.

Here’s another action the USSF should take to boost homegrown talent: actively identify and track goalkeeper development across the country with a data-based analytics package like Stopper. Rolling out Stopper to elite player development programs across the country – say U-13 and up – would instantly provide the USSF with an up-to-the-minute performance overview of the nation’s most talented up-and-coming GKs and also offer a window into their long-term development.

Data means nothing if it’s not actionable, so imagine implementing Stopper in combination with hiring a national Director of Goalkeeping. Suddenly, goalkeeper development in America can be held to a measurable standard with oversight at the highest levels, and areas of deficit can be addressed at a national level in real time. Player performance and development can be measured and assessed objectively from the earliest ages all the way to college and beyond. Which in turn would make it easier to support and develop coaches, and hopefully provide a pathway for growing the overall pool of USSF-licensed GK coaches.

As for the CHL foreign goalkeeper ban? It was lifted this summer amid mixed results. Bottom line: as Bill Reno notes, the best solution to a lack of goalkeepers in the development pipeline is not to shut out competition – it’s to build a better pipeline.

Goalkeeper Actions Per Game

By on November 12, 2018 in Uncategorized

When we started Stopper, our assumption was that tracking a goalkeeper manually is much easier than tracking an outfield player. Why? For one, their positioning is generally predictable; for another, they tend to spend less time on the ball then other players.

But is that assumption true? How many times does a goalkeeper really touch the ball in a game?

Turns out it’s surprisingly hard to find the answer. Among the best-known analysis is a study published in the Journal of Sports Science that tracked the movement patterns of 30 French League 1 matches over a period of two seasons. Among other findings, it showed that players had, on average, 47 possessions per match and 2 touches per possession. Central forwards had the fewest possessions (35) while outside defenders had the most (56) – but I couldn’t find a reference to goalkeeper possession per se.

A more recent study published in PLoSOne showed that while goalkeepers and centre forwards had the fewest touches on the ball, GKs had the longest “In Ball Control” intervals and the highest proportion of “In Ball Action” intervals with control – likely because they are allowed to catch the ball and can’t be attacked while they are holding it. Again, while informative, this study didn’t specifically track the amount of goalkeeper touches per game.

OptaStat is always a good source for detailed football data for the big leagues in soccer – doing a quick assessment of the past season, it appears that goalkeepers in the EPL average around 44 touches per game if we include the “accurate long balls” category. I suspect that is a subset of Opta’s goalkeeper passing and goal kick data, in which case the average amount of GK actions, as tracked by Opta, drop to around 37 or so per game, roughly that of a striker.

Our own Stopper data is obviously entirely focused on goalkeeper performance – and since touches per game is something that we do track (as a byproduct, not a specific data set) we thought it would be interesting to look at the numbers. With a couple hundred games of testing in the vault (both MLS and MLS Academy games, so not a clean data set), the official Stopper number for average GK touches per game clocks in at a surprisingly high 47.26 actions per game. The highest number of actions recorded in a single game was 76; the lowest on record is 32, variables that depend on style of play, opponent, defensive strength and so on.

I’m going to be keep an eye on this stat – is there a connection with wins and losses? Do keepers rated higher in Stopper also touch the ball more often? – but the bottom line is goalkeepers seem to get their hands, feet and bodies on the ball a surprising amount these days, making the #1 truly the 11th man in modern soccer!

Database Issue Resolved

By on November 1, 2018 in Uncategorized

UPDATE Nov. 2/18: the Stopper database is up and running again. Please let us know if you experience any issues.

Nov. 1/18: Our database hosting service is down – we’re working to resolve the issue, but in the meantime you won’t be able to record games or access game data. We’ll post an update here shortly. Sorry for the inconvenience!

18,565 touches of greatness

By on October 3, 2018 in Uncategorized

We’re proud to announce that as of today Stopper has recorded a total of 18,565 goalkeeper actions: 4059 Saves, 1044 Goals, 1043 Crosses, 8647 Distributions and 3772 Communication. How cool is that??! We’ve got hundreds of users tracking goalkeeper stats in countries as diverse as the US, the UK, Norway, Australia, India, Mexico, Russia and Greece.

Considering that football/soccer is the world’s most popular game, goalkeeping can be a bit of a lonely job given that it’s a specialized position and many clubs don’t give it the attention and support it deserves. When we launched Stopper, we couldn’t imagine that it would become so popular worldwide – good to see so many people out there supporting GK development.

Thanks to all our users for their support and all their suggestions for making Stopper better. Look forward to reaching even more #1s and celebrating their greatness! #gkunion

NE Revolution Partnership Continues

By on August 30, 2018 in Uncategorized

Since 2017, we’ve had the pleasure of working with the Goalkeeper coaches of the New England Revolution Youth Teams. As part of a longer-term pilot program, we’ve been working with the staff to track goalkeepers in the U13 and U14 program and incorporating Stopper game data into the Academy’s high-performance keeper training program.

According to Karl Spratt, Head of Academy Goalkeeping, “Stopper give us another avenue to really drill down into specific aspects of our goalkeepers’ performance when we review games, and can then look at how they’ve developed over the course of an entire season. This data then provides us with a tangible benchmark for our goalkeepers that is easy for them to visualize.”

The New England Revolution Academy is one of the top youth development environments in the US Soccer Development Academy system, as evidenced by top tier home grown talent like Eliot Jones – recently called up to the USMT U16 Team in April 2018. Having their insight and feedback on Stopper has been invaluable, allowing us to continually improve the software and build in new features specifically aimed at high-performance goalkeeper training.

You’ll never walk alone

By on June 15, 2018 in Uncategorized

It was never going to be an easy game. But with a measure of luck behind the offensive genius of Mo Saleh, there was a betting man’s chance LFC would beat Real to finally hoist the Champions League trophy for the first time in a generation.

Then at the 31st minute, it all fell apart.

Goalkeepers don’t win games but they sure can lose them. That’s why it’s the hardest position in the beautiful game – all the guts and rarely the glory. Strikers are allowed plenty of mistakes; mids and defenders maybe one; for goalkeepers, mistakes are fatal and unfortunately Loris Karius made not just one, but two.

Our mission at Stopper is to help goalkeepers and coaches use in-depth performance data and analytics to understand their strengths and weaknesses. One of the key metrics we measure is decision making around distribution, by tracking completion percentages across a range of actions. Based on the data from the goalkeepers we’ve analyzed to date, I can almost guarantee that if we tracked Karius’ game performance and development over the years, a pattern of making distribution mistakes in high pressure scenarios would appear. I’m pretty sure that Stopper would also show a pattern of compounding errors – where a first mistake leads to a loss of composure that almost inevitably leads to yet more mistakes.

Moral of the story: the more objectively coaches can assess goalkeeper performance the better. By integrating Stopper into their high-performance training, teams can identify player performance trends and subsequently make the appropriate coaching decisions. When the Dutch team swapped keepers in the dying minutes of their game against Costa Rica in the 2014 World Cup, it was lauded as a bold move: that brilliant coaching call by van Gaal was exactly the kind of decision-making Stopper is designed to support.

Let’s say that Jürgen Klopp had our analytics at his fingertips and knew that once Karius had made his initial mistake, there was a very high-percentage chance of getting scored on again due to loss of composure. Then the decision would have been a no-brainer – sure, a tough call, but imagine for a second that based on that data, he pulled a play from the hockey book. Imagine he had substituted Mignolet or maybe Bogdan for Karius – and then imagine that subsequently Bale’s shot in the 83′ minute never went in.

Football is a game of what-ifs and might-have-beens, but if Klopp had the data to support that call, there’s an even chance that the Champions League trophy would be standing tall at Anfield today.