I got really sick of just reading those quick match previews that barely scratch the surface. You know the drill, “Atleti are strong at home,” “Celta need points.” Big deal. I wanted to see the guts of the numbers for this specific matchup, not just general vibes. So, I decided last Thursday, instead of watching old games, I was going to pull the stats myself and see what the real story was. This wasn’t some fancy data science project; this was me, a spreadsheet, and a lot of copying and pasting.
The Initial Data Hunt – A Real Mess
My first step was pure chaos. I decided I needed three main buckets of information: historical head-to-head metrics, recent team form (the last five games), and key player individual stats. I started hitting up the usual suspects for football data. Trying to find clean, easily downloadable numbers for just the Atleti vs. Celta matchup over the last few seasons? Forget about it. They make you work for it.
I ended up with about six browser tabs open. I manually copied over the total number of shots, shots on target, and—the big one—Expected Goals (xG) for both teams in their recent outings. Trying to match up the xG data from one source with the possession numbers from another was a total nightmare. I spent a solid hour just wrestling with my Excel sheet, making sure “Atlético Madrid” wasn’t accidentally entered as “ATM” in one column and messing up the averages.
I focused on these key figures because they tell a better story than just the scoreline:
- Average Defensive Actions per 90 minutes (Atleti is usually great here, but I wanted to see if Celta could pressure them).
- Possession percentage away from home (Celta’s Achilles’ heel).
- Set Piece Goals conceded/scored in the last 10 games (A massive potential difference-maker).
Crunching the Team Form – Finding the Dip
Once I had the raw numbers shoved into the sheet, I started running the averages. I quickly realized that simply averaging the last five games didn’t tell the full story. Both teams had played easier fixtures mixed with heavy hitters. I had to weight the scores. I decided to give the most recent three matches 60% of the weight, and the previous two only 40%.
What popped out immediately? Celta’s inability to maintain defensive structure after the 70th minute when playing away. I calculated their total number of conceded shots between minute 70 and 90, and the numbers spiked up like crazy compared to the first half. That’s a key piece of information for Atleti, who often score late.
I also filtered the historical data to only include games where Atleti had Simeone managing, because anything before that is irrelevant noise. It’s a completely different team. This simple filter immediately showed Atleti’s dominant home record against Celta, often winning by a single goal but maintaining control.
The Player Form Breakdown – Diving into the Individuals
This was the most fun part, but also the most tedious, because player ratings and stats are often scattered everywhere. I decided I needed to look beyond Griezmann’s goals and Aspas’s assists and focus on specific metrics related to their roles.
I tracked the successful dribbles and tackles won for Marcos Llorente and compared them to Celta’s central midfielders. The result was clear: Llorente had been a beast, especially in transition. That confirmed my suspicion that the midfield battle would be tilted heavily in Atleti’s favor, assuming Llorente started wide right or central.
Then there’s the striker comparison. I pulled the xG numbers for Morata and compared his conversion rate over the last month to his season average. His recent form was actually slightly depressed compared to his overall stellar season, meaning he was due for a goal. This is the kind of detail you miss when you only look at the match report score.
I also highlighted the passing percentage for defenders—specifically Witsel for Atleti. He was maintaining a ridiculously high pass completion rate, which suggests Atleti’s build-up play remains stable and reliable, no matter who they play.
It took me maybe five hours total, sitting there moving numbers around and arguing with my spreadsheet, but I ended up with a far better picture than any two-minute news clip could give me. This whole process confirmed that if you want the real story, you have to get your hands dirty and manually stitch those disparate stats together. Sharing this because maybe it saves someone else the pain of the initial copying phase!