The Thomas Rongen Line
/How good or bad is a soccer coach? For most of the world, it’s a hard metric to determine. Because the squads for teams vary so highly based on the amount of money the team can spend (e.g., Manchester City alone has a roster valued, conservatively, more than the entire rest of the Football League combined), it can be difficult to determine if a manager is successful due to their efforts, or due to a hyper-talented team. Points Per Game (PPG) is not adequate.
I don’t think anyone would consider Avram Grant a great manager, but he did manage to get an expensive Chelsea squad to the 2008 Champions League Final. So can we then calculate Points Per Game Per Currency Unit (PPGPCU)? That also gets complicated. While the Euro theoretically equalizes currency values across most of Europe, that doesn’t equate to the wildly varying values of the world’s other 179 currencies. We could use an individual nation’s currency (e.g., the US dollar) - and then value all world currencies against it and try to normalize that way, for Points Per Game Per Normalized Currency Unit (PPGPNCPU).
However, we then run into the problem of relative resources per league. The top five European leagues have vastly more money in them overall than, say, the Estonian A. Le Coq Premium Liiga. Also, the quality of professional leagues varies significantly. Many leagues have one Big Club and a bunch of smaller hopefuls - witness the historical success and resources of Ferencváros versus the rest of Hungarian soccer. Thus PPGPNPCU would be vastly skewed in favor of a manager in a smaller, less competitive league with less money in it overall. While you could possibly normalize resources within a given league, it becomes very difficult to compare managers between leagues.
Okay, so how do we fix that? We could try to use one of the various methods to rank relative league strength - Opta publishes ELO data on this - and then generate a Points Per Game Per Normalized Currency Unit Normalized By Relative League Strength (PPGPNCPUNBRLS). The problem is that the ELO is skewed because, for instance, English Championship teams rarely, if ever, play Liga MX teams in competitive matches. The sample size is too small. While I trust that Opta has done a ton of work on this, I’m not sure how you get past this hurdle. To my mind, at least, this is where the increasingly-unwieldily-named PPG metric falls short of being truly useful.
Many smarter minds than I have developed metrics to try to answer this question, with, I’d argue, limited success.
But - what if you had a league where the money spent per team is equal (at least nominally) and the currency is the same (or at least heavily normalized between fixed member nations)? We have that in Major League Soccer. MLS also has the great equalizer of the draft, which (hopefully) gives bad teams a chance to quickly get back on their feet. Given the normalized resources per team, the results can be much more confidently attributed to the manager. Here we can realistically use the basic Points Per Game metric and be quite confident that it’s an actual indicator of managerial quality.
I was lucky enough to find that John Lenard compiled a dataset of all MLS coaching records (for only MLS games - no Cup competitions) through the end of the 2018 season. All I needed to do was update it through 2024. So, through Apr 24, 2024, I present to you MLSCoachStats. I’m working on finding an inexpensive enough API data provider to write some Python to auto-update it, but I’m not there yet. There you’ll find the records and PPG for every coach to have ever stood on the sidelines, including Roy Wegerle, interim manager for Colorado for one game. As a side note here - I support Austin FC and Blackburn Rovers. I am sad to report that every coach for Austin, and every former Rovers player who has managed (Wegerle, Nelsen, and Friedel) have been quite sub-par.
The two best managers, as judged by PPG with over 100 games? Bruce Arena and Octavio Zambrano. Bruce is slightly above Zambrano at 1.658 PPG to 1.657, and gets the heavy credit by virtue of his 401 more games coached. Only Sigi Schmid, at 547 games, tops Arena’s 544. In active managers it’s Wilfried Nancy. Arena has been cleared by MLS to return to coaching, so he may regain this title soon.
How do you determine the worst, or, at the very least, the MLS coaching equivalent of the Mendoza Line? The Mendoza Line is baseball’s “shorthand for offensive futility” - having a batting average under .200. It’s generally used as the easy litmus test for whether a player is definitively bad. The best equivalent metric for that is the average minimum PPG to make the MLS playoffs. In talking with the amazing Phil West, journalist extraordinaire, he suggested the Thomas Rongen Line. Rongen, over a career with four different MLS clubs, chalked up a 1.32 PPG lifetime average. In a 34 game season, 1.32 PPG gets you to ~45 points which is right in the mix for the playoffs. The average PPG required to make the playoffs across all MLS seasons is 1.29. I’m happy to report Rongen (like Mendoza before him) is over his eponymous line.
If you’re a pedant (and look, you’re reading a soccer analytics site - you know you are), you might argue that the early seasons of MLS skew this number. In 1999 the PPG required to make the playoffs was an astounding 0.91. The Modified Rongen Line comes out at 1.37 PPG if you only count 34 game seasons.
So who is the worst of all time with over 100 games? Barely eking it out at 1.1765 PPG is the (now) immortal Veljko Paunović. Hoist a crisp Serbian Jelen in his honor as you contemplate the horrorshow that was the 2015-2019 Chicago Fire (it would appear he was so bad that the ratings for the TV show Chicago Fire were down until after he left). Then you can marvel at how he managed to get a job at Reading, get 1.24 PPG, and then somehow get a job at Chivas. He is truly a statistical outlier of the highest caliber.