There'll be another terrific crowd at The Racecourse for the visit of Halifax, which means it's time to address a common concern among supporters.
It's a complaint I've heard Wrexham fans make ever since I started going to The Racecourse (and believe me, that's a long time ago): do Wrexham lose their nerve in front of a big crowd?
We heard that query a lot in the first half of the season. We all turned into amateur psychologists, analysing whether the difference between our performances at home and on the road were due to an inability to perform in front of a big crowd.
I'll declare my interest now: I really didn't think that was the case, but when we started discussing the matter on "DragonHeart" on Calon FM, I noticed the amount of reaction the subject provoked and thought it was time to investigate the idea. The video below will start at the beginning of a discussion of my findings.
The notion that we don't perform in front of large crowds is based on gut-feeling and confirmation bias. We feel like it's true, and then we go looking for evidence to confirm our hunch, ignoring evidence that doesn't fit into our theory.
I needed something more concrete than that, but simply working out the average points we win per game and comparing that figure to our average result in front of a variety of attendance sizes doesn't cut it. I wanted a second level of data to compare the results to, because such a simple analysis of results alone would lead to the findings becoming unreliable.
For example, if we beat a side that hadn't won a point in 25 games we'd only be doing what we'd be expected to do, yet when we compare the 3 points we earned from that game against the average points we win per game, it'd appear to be an over-performance.
I needed a better way to measure our results, but first I needed to make some decisions about the data I’d use. Obviously, I only used home games, because I wanted to compare like-for-like statistics, and didn’t want outliers like the sub-four figure crowds we sometimes experience in away games.
For the same reason, I decided to go back to our first season in the National League, as that would mean all results would be achieved at a consistent level of competition. I also opted to only use league and play-off matches, because that would rule out the extremes of attendance and performance which cup ties bring.
For instance, if we included the Gloucester City game from last December, it would encourage the idea that we perform well in front of small crowds: the attendance was very small, and we won 5-0.
In actual fact, we played a side from the bottom end of the league below us, who had half their players out with COVID, and selected a team which contained a number of youth players making their debuts. Meanwhile, our side was very strong, apart from the absence of Paul Mullin. Considering all that, I’d say 5-0 was pretty much par for the course.
The opposite is true as well: say we faced Manchester City in the FA Cup and drew 0-0 in front of a capacity crowd: clearly that would be a massive over-performance. But without context, it’s just a home draw in which we failed to score: it would rate the same as our draw with Wealdstone earlier this season! Across all home games in the National League we average 1.83 points per match, so a draw with City would represent an underperformance of 0.83 points against our usual result!
So, to measure whether we’re performing above or below expectations more accurately, I looked up the final league position of each team, and worked out the average points per game we’ve won against every position. For example, when we play teams which end the season as champions we’ve won an average of 1.7 points per match, while when we entertain the sides that finish bottom we average 2.64.
The table above shows our average points haul from a home game in the National League against every opponent's final position. To encorporate this season's results, I used our opponents position in the table at the start of the game.
Now I had to use this data to draw some conclusions. The basis of my calculations was that if we beat the side which finishes bottom, we've overperformed by .36 points compared to how we usually fare against such teams; if we draw, we're 1.64 points below expectations. By taking these figures and comparing them against our results in front of a variety of crowds, we'd be able to make some judgements. Whether we win, lose or draw in front of big attendances is too simple; whether we perform above or below expectations, taking the quality of the opposition into account, is much more revealing.
Next, I split our matches into groups according to attendance: a new group formed every 500 spectators. That allowed me to see if we got the results we'd expect if the crowd is bigger and lower.
The chart below displays my findings: if the bar is above the x axis, we overperform in those matches; if it's below the axis, we underperform.
I should add a word of warning. Clearly results will be mre reliable if there is more data to analyse. Therefore, we should ignore the first column, which suggests we underperform when the attendance is above 9,500. The truth is, we've only attracted such a crowd once to The Racecourse in the National League, so the 1-1 draw with Torquay is the only result in that category.
The pattern seems to suggest that, if anything, large crowds inspire us. Ignoring that Torquay outlier, if a crowd is above 7,500 we generally do better then expected. Attendances in excess of 9,000 appear to inspire us, as we really overperform massively in those games, more than in any other attendance category, virtually gaining a point extra compared to what we ought to win.
Indeed, the greatest extremes come when the crowds are high. When the attendances are smaller, our performances seem to be much closer to what you'd expect. While we have struggled when the crowd has been betweem 6,000 and 7,000, we've done spectularly better than expected when the crowds are bigger than that. This should seem to suggest that there is a pattern here: big crowds certainly affect our performance. However, far from the pessimistic view of the team freezing on the big occasion, it seems that they are inspired by the high expectations.
In front of small crowds we seem to potter along, pretty much doing what's expected of us; in fact, if anything we do slightly worse than expected in those games.
All studies of this kind require an element of caution. There's a chicken and egg scenario to consider: do we perform in front of big crowds because they inspire us, or is a big crowd attracted because our team is strong, in form and therefore more likely to play well? At the other end of the scale, do we not stand out in front of smaller crowds because our team is weaker, which is why the crowd was small in the first place?
There's also a question in my mind about the reliability of the figures at the extreme ends of over- and under-performance: Three points for a win and one for a draw perhaps skews the figures as you're rewarded disproportionately for winning. However, that's a calculation my A Level in Maths hasn't equipped me for!
The broad conclusion is sound though: we don't underperform in front of big Racecourse crowds, and in fact we appear to relish the motivation of a large, loud backing. So remember that in our remaining home games of the season: the more of us that are there, and the more nose we make, the better!