Playing a game under Opta's watch: heat maps, pass completion and terror

SUTTON, England -- Many of us involved in football reporting have, at some stage, been told our opinions are irrelevant because "we've never played the game." That is, of course, entirely untrue. We've all played the game; we're simply not very good at it. Last week, some of us were afforded the pleasure of discovering quite how incompetent we were, with statistics, graphics and heat maps to prove it.

Football data company Opta held a friendly match at Gander Green Lane, the home of Sutton United, who hit the headlines following their FA Cup defeat to Arsenal last season mainly because reserve goalkeeper Wayne Shaw was filmed eating a pie in the dugout.

Our team features journalists, television producers and others who work with Opta's data, some of whom boast fitness levels sadly comparable to Shaw's. Our performances were analysed in the same forensic manner of Premier League matches too, a somewhat daunting prospect for those of us who hadn't played 11-a-side for years. Everything -- passes, shots, tackles, dribbles, fouls -- would be tracked.

I played in the No. 7 shirt as the right-sided midfielder for Opta Blue, who ran out comfortable 8-3 winners over a disjointed Opta White side. But everyone's main focus, of course, was discovering their individual statistics after the match.

When you realise you're being tracked statistically, you start the game very conscious of your actions. I literally didn't get a touch of the ball in the first five minutes, simply shuttling up and down on the right (see my heat map) and shouting for the ball while the central midfielders got stuck in. Usually, no one would notice how little I was involved; today, my lack of contribution would be displayed for all to see.

Things changed after five minutes. My first involvement was collecting a loose ball inside my own half. Determined to get my pass-completion stats going, I knocked a safe 10-yard pass. The good news, however, is that this prompted a flowing passing move down the left of the pitch, and as I continued my run, drifting inside into a position between the lines, I found myself receiving the ball 25 yards out.

Keen not to embarrass myself by using my left foot, I channeled my inner Ricardo Quaresma, struck the ball with the outside of my right foot and was delighted to see it bend, bounce awkwardly in front of the goalkeeper and nestle in the bottom corner: 1-0.

My opening goal, I later learn, had an "expected goals" (xG) value of 0.03. "Expected goals" is increasingly used among football statisticians to summarise how good the sides' chances have been. Taking the position from which the shot is attempted, the nature of how the ball arrived at the shooter and cross-referencing it with several thousand shots from a statistical database, it essentially works out the chance of scoring in that situation. A value of zero means there's no chance of a goal; one means a certain goal. Therefore 0.03 is, depending upon your perspective, either a poor decision to shoot or a fine finish.

"Expected goals" is a genuinely insightful and increasingly prominent part of football data, although it has faced some criticism from ex-professionals, with Craig Burley memorably slamming the concept on ESPN FC. A personal belief is that the metric is excellent. We all know that simple "shots" and "shots on target" figures don't tell the whole story, as a shot on target from five yards isn't as valuable as one from 50. So why not statistically calculate the probability of each shot ending in a goal? It's as simple as that.

"Expected goals" is a peculiarly unsatisfactory term, however, and any number expressed as a decimal starting with zero instinctively feels unattractive. A term like "shot value" probably makes more sense, and for individual shots, a percentage could be used instead. "I attempted a 3 percent shot" feels much more appealing than "I attempted a 0.03 shot."

Anyway, the first half goes reasonably well and players from both sides congratulate me on that opening strike as we head for the dressing rooms. We're 6-1 up, I've scored the best goal of the game and played possibly the best pass too -- what I'd like to think was a Pirlo-esque diagonal in the lead-up to our fourth.

Yet upon glancing at the live stats on an iPad, to my horror I realise I've recorded a pass completion rate of just 30 percent. All those poor touches in midfield that concede possession but are quickly forgotten at Sunday League level are no longer forgotten when Opta are on your case. My seven misplaced passes are presented with huge red arrows all over the iPad screen. I have the worst pass-completion rate on my team, and one of them -- when I was caught out while briefly playing at centre-back -- directly cost us our only concession.

My opener no longer matters. The fact we're 6-1 up no longer matters. I'm completely focused on my passing figures.

My second half-performance, therefore, is the most blatant display of statistic-massaging Opta will find in their extensive database. An outsider might witness my performance and suggest I'm being tactical, seeking to preserve our five-goal lead and shut the game down by playing cautiously and keeping possession. Oh no. Instead, I'm trying to improve my passing figures up, fearing the postgame debrief.

My timeline is simple. A safe sideways pass in the 46th minute. A safe sideways pass in the 51st minute. In the 55th minute, I try to play a more ambitious ball and concede possession, so then, in the 60th minute, I play another safe sideways pass.

Things become particularly desperate in the 76th minute after we've conceded again, as I take advantage of our striker being substituted by trotting across to take the kick-off, therefore ensuring another complete pass.

By this stage, I'm exhausted and ready to depart, so my final pass is a Hollywood ball, a 40-yard chip over the defence to create a one-on-one chance for our centre-forward, who forces a fine save from the goalkeeper. This adds a pleasing yellow "chance created" arrow to my otherwise underwhelming chalkboard and mercifully ensures I'm over the 50 percent pass-completion rate.

What the statistics and graphics captures perfectly is not simply my overall ability but also my style. I'd love to be an amateur version of Michael Carrick: a consistent, solid, positionally disciplined midfielder. Instead, I'm more of an amateur Luis Garcia, flitting around on the edge of games, providing a couple of moments of magic to look good in the highlights, but infuriatingly inconsistent and liable to conceding possession.

Indeed, in my personal postgame analysis, I'm far more concerned by my errors being highlighted than my good moments being visible.

In a professional setting, of course, this data would be analysed by my coach and I'd be told that my passing figures were very low. It would also be analysed by the opposition coach for my upcoming game, who would seek to exploit my weaknesses. Therefore, were I playing in a second leg of this contest, I would be overwhelmingly focused upon keeping things simple and not repeating the same errors, perhaps at the expense of the ambitious passes and shots that proved successful.

I tend to think that statistical analysis encourages this, as players seem to focus on improving their weaknesses, rather than concentrating upon their strengths. Indeed, the recent development of football has been all about universality: players becoming more similar as all-rounders, rather than specialists. Defenders must be comfortable at starting attacks, attackers must start the defensive pressure and underperformance in any area will be scrutinised by the likes of me. Or, perhaps, by themselves.

Poring over your own mistakes can be difficult, but even Premier League players benefit from being able to read, interpret and analyse data personally.

Meanwhile, I'll take solace in the fact that even Cristiano Ronaldo and Lionel Messi won't be able to beat my 100 percent shot-conversion rate for the season.