What Pf And Pa Actually Measure In Football Data
Pf and Pa are not the kind of stats you find on a broadcast graphic or a basic match report. They come from event-level data providers like Wyscout, Infotrack, or StatsBomb, and they measure progressive ball movement. Pf, or progressive carries, tracks every time a player dribbles the ball forward across a yardage threshold that meaningfully advances the attacking position. Pa, or progressive actions, is broader — it includes progressive passes and progressive carries combined into a single metric that captures forward-moving contributions regardless of whether the player has the ball on their feet or is playing it. I spent years watching highlight reels before I started working with actual event data. The shift was jarring. A midfielder who looked invisible on TV would consistently rank in the 90th percentile for Pa because his distribution repeatedly breaks lines. A winger who looks spectacular might barely register because most of his carries go sideways or backward under pressure. These stats don't care about aesthetics. They care about net forward progress. The methodology behind Pf and Pa isn't secret, but the exact thresholds vary by provider. Wyscout defines a progressive carry as one that moves the ball at least 10 yards toward the opponent's goal. Pa is calculated by counting every pass or carry that crosses a defensive line or moves the ball into the final third. Some providers use spatial heat zones, others use yardage, and a few combine both. If you're building your own model or comparing across platforms, you need to know which definition you're working with. Mixing them is the fastest way to get nonsense results.
I ran into a concrete problem early in my career that took me three weeks to fix. I was cross-referencing Pf data from two different providers for the same season across La Liga. Player A had a Pf rate of 4.2 in one dataset and 2.1 in the other. Both claimed to be measuring progressive carries. The discrepancy wasn't a bug. Provider one used a simpler threshold of 5 yards for any forward move. Provider two required a minimum of 10 yards and also excluded carries that started in the defensive third and only crossed the halfway line. I was essentially comparing two different measurements and wondering why the rankings looked nothing alike. The workaround was straightforward: I wrote a normalization script that converted everything to a common baseline using yardage thresholds and only counted actions originating in the middle or final third. After that, the data aligned and the analysis became usable.
Why These Metrics Matter More Than Expected
The reason Pf and Pa exist is simple. Traditional stats like total passes, tackles, or distance covered tell you volume. They don't tell you direction or purpose. A player can complete 80 passes and contribute nothing forward. A player can attempt 6 progressive carries and shift the entire balance of a match. Pf and Pa cut through the noise. Here's something that surprised me when I first started paying attention. Pf correlates more strongly with expected goals created over a season than a player's total distance covered or even their pass completion rate. There's a reason for that. Carries into advanced areas force defensive reactions. They compress the opposition shape. They create passing lanes that wouldn't exist otherwise. Pa captures that effect at a higher level because it includes the distribution side of progressive play. A player who receives the ball under pressure and immediately plays a line-breaking pass is generating value that a stat line full of safe lateral passes never shows. I once worked with a coaching staff that was frustrated because their star attacking midfielder looked inefficient on paper. His pass completion was below the team average, his dribble success rate was middling, and his total shots were low. When we pulled his Pa per 90, it was off the charts — top 3 percent in the league. He was failing on safe passes because he was choosing progressive options that carried risk. His numbers on traditional metrics looked bad because the metrics didn't account for decision quality under pressure. We adjusted our evaluation framework to weight Pa andPf heavily and stopped using raw pass completion as a standalone filter. It took two transfer windows for the staff to fully buy in, but the data held up.
Get the Full Details
How To Read These Numbers Without Getting Misleading Signals
Rates matter more than totals. A fullback who plays 90 minutes in a possession-dominant system will accumulate more Pf and Pa than a winger in a counter-attacking side, even if the winger is more dangerous in transition. Always normalize to per 90 or per 100 possessions. I prefer per 90 because it accounts for playing time variations, but per 100 touches gives you a clearer picture of decision-making efficiency when comparing players with very different workload levels. Context is everything. Pf and Pa are meaningless without knowing what a team's tactical structure looks like. A team that plays through the middle will have a different profile than one that switches play repeatedly. Wide teams will show high Pa from fullbacks and wingers but lower central Pf. Midfield-centric teams will concentrate their progressive numbers in central zones. If you're scouting or building profiles, layer in spatial data. See where the progressive actions are coming from and where they land. A high Pa number from a player whose progressive passes all go sideways into dead zones is not the same as a high Pa number from someone consistently finding space between the lines. I encountered a borderline case that nearly made me discard Pa entirely for a season. A center-back in a top European league was ranking above several elite attacking midfielders in Pa per 90. At first glance, this looked like the metric was broken. It wasn't. The center-back was playing a very direct style with long progressive passes that broke the first defensive line. These passes traveled 30 or 40 yards and landed in advanced areas. The attacking midfielders, by contrast, were operating in tight spaces with short progressive passes and carries that still broke lines but covered less ground. The old instinct was to dismiss the center-back's numbers as a fluke of length over quality. I pushed back and pulled the end locations of each progressive pass. About 40 percent of his progressive passes ended in the penalty area or the immediate surrounding zone. The rest went to wingers in advanced positions. The metric wasn't broken. It was just measuring something broader than I expected, and I had to adjust my interpretive framework rather than throw the data out.
The Practical Application
If you want to use Pf and Pa in your own work, start simple. Pull per 90 rates for your target league and position group. Filter by minimum minutes played — anything under 800 minutes in a season skews heavily. Rank within position groups rather than across all players because roles differ too much. A center-back with 2.5 progressive actions per 90 is elite. A central midfielder with 2.5 is average. The benchmarks shift entirely based on position and tactical role. Combine Pf and Pa with other contextual metrics. Progressive passes completed percentage tells you reliability. Progressive carries into the penalty area tell you threat level. Shots from progressive entries tell you finishing output after breaking lines. Defensive actions lost during progressive carries matter too. A player who loses the ball immediately after a successful progressive carry is generating false positive value. I've found that the most useful application of these stats is in identifying undervalued players. Scouting departments and analytics teams often focus on visible metrics like goals and assists. Pf and Pa highlight players who are creating forward momentum without the final product showing up in traditional box scores. These are the players teams should be monitoring before markets price them in. I've seen clubs sign midfielders purely on Pa trends before the broader scouting community caught up. The signal appears six to eight months before the narrative shifts.
When These Metrics Fail You
Pf and Pa are not universally applicable. Set-piece heavy teams will inflate Pa through direct throws and corners that don't reflect open-play progressive structure. Teams that play long balls into channels will look dominant in Pa but generate low expected threat because the progressive action ends in a loose ball contest rather than a structured chance. Transition teams are another edge case. A team that wins the ball in its own half and immediately plays one long forward pass will rack up Pa numbers that don't predict sustained attacking performance. These metrics also struggle in low-scoring, low-possession environments. If a team rarely enters the final third, both Pf and Pa will be artificially suppressed across the roster. That doesn't mean the players are worthless. It means the structural context of the team's play style limits what these particular stats can reveal. In those cases, you need to pair Pf and Pa with metrics like progressive passes allowed (how often a team forces opponents into non-progressive situations) or defensive line-breaking actions to get the full picture. My recommendation when dealing with these limitations is to build a composite index rather than relying on a single metric. Combine Pa per 90 with progressive carry success rate, final-third entry frequency, and expected threat generated from progressive entries. No single number tells the whole story. The composite does, and it takes about twenty minutes to set up once you have the raw data pipeline in place.
What To Do Next
Start by getting access to event-level data. Most professional providers offer sample datasets. Match the Pf and Pa definitions against your existing framework before adopting them. Run a season-long validation test on a position group you already understand well. If the rankings make intuitive sense alongside what you see on the pitch, you're using the right thresholds. If they feel backwards, adjust your definition or switch providers. I still make mistakes with these metrics. I once flagged a player as elite in Pf before realizing his progressive carries were happening mostly in wide areas where they had minimal impact on the central attacking structure. The raw number was fine. The context was missing from my initial read. That's the ongoing lesson with Pf and Pa. They're powerful tools. They're not complete answers on their own.
