Introducing @CzechFooty’s 2022/23 Fortuna:Liga Team Preview Series
Last year was a hot-needle test run. This year is the showtime. That’s right: we go bigger and, hopefully, better on comprehensive Fortuna:Liga team previews that made a premiere in this space about 12 months ago. Below is the 2022 roadmap coupled with an obligatory “what to expect”.
The roadmap
Unlike last year, I have more time to plan it all out and, as a result, spread it all out, too. Going by the 2021 feedback (but also just plain common sense), three articles per day are not easy to swallow for any regular reader who has a life besides chewing on the products of my compulsive writing. This year, therefore, there’ll be a maximum of two per day (morning-evening) while the popular contenders along with the last cup participant get their unique slots. Besides, there are two little breaks sprinkled inbetween for both me and you.
Just like I did a year ago, meanwhile, I’ll try to pair regional rivals together when possible. Pardubice and Hradec Králové shall go hand-in-hand for the first time, and so will Baník Ostrava and Zbrojovka Brno as the best-supported representatives of Silesia and Moravia historical regions. The Jablonec-Liberec pairing can’t miss out either, of course. Bohemians and Mladá Boleslav rent out their stadiums to open 2022/23 once again, so they deserve to go together, while Trinity Zlín and Sigma Olomouc each have one of 14 regions named after themselves and stand alone as two clubs with sponsors/partners in their names, so there you go — connections everywhere!
Saturday 16 July: FC Viktoria Plzeň (season opener on 19 July)
Sunday 17 July: SK Slavia Praha (21 July)
Monday 18 July: AC Sparta Praha (21 July)
Tuesday 19 July: 1. FC Slovácko (30–31 July)
Friday 22 July: SK Dynamo Č. Budějovice (30–31 July)
Saturday 23 July: FK Mladá Boleslav (30–31 July)
Saturday 23 July: Bohemians Praha 1905 (30–31 July)
Sunday 24 July: FK Jablonec (30–31 July)
Sunday 24 July: FC Slovan Liberec (30–31 July)
Tuesday 26 July: FC Hradec Králové (30–31 July)
Tuesday 26 July: FK Pardubice (30–31 July)
Wednesday 27 July: FC Zbrojovka Brno (30–31 July)
Wednesday 27 July: FC Baník Ostrava (30–31 July)
Thursday 28 July: FC Trinity Zlín (30–31 July)
Thursday 28 July: SK Sigma Olomouc (30–31 July)
Friday 29 July: FK Teplice (30–31 July)
What’s new in stock
Expanded starting field: As you can see above, I won’t be skipping the promoted side this time around. Last year, I had a good excuse since Hradec had last been part of the top flight in 2017 and I didn’t really know any fans to speak to, but this year I couldn’t justify it anymore. I actually have a whole 2020/21 dataset on the remaining few Zbrojovka players and plenty of fans to survey. That said, the Brno team preview will naturally look different — mostly relying on the numerous fans recalling the promotion campaign and assessing the summer full of reinforcements and deadwood shedding.
Preview lay-out: Each piece is once again divided to 3 “time zones” — looking back on 2021/22, taking a temperature check of the off-season, and looking ahead to 2022/23. There’ll be subtle tweaks along the way: the off-season section now features a “need left to be addressed” category (one slot I’d expect them to fill eventually, or else they’re begging for trouble) and the original “biggest addition/greatest subtraction” twin section morphs ever so slightly into “biggest upgrade/downgrade”. After I published a two-part article based on my MVP model, the “most valuable player” section as part of each team preview is now to be my subjective call (supported by data) rather than simply what the model chucks out like in last year’s case. One category — “statistical trend to follow“— is gone, but that’s only to make way for…
Season forecast: The one innovation I’m by far the most proud of. I partner with the excellent Jakub Černý of @ČeskýMistr to bring you the inaugural prediction of how the next season shall unfold. It’s not going to be your typical off-the-cuff “14th”, but rather a multi-layered result of dozens — if not hundreds — of simulations based on last term’s efforts and off-season moves. I’m also pleased to say Jakub has come up with another The Athletic-inspired graphic, so the team previews shall be even more colourful and intriguing.
A new-look comparison tool: These tables will no longer appear anywhere:
It’s not like we received any negative feedback or that Adam wouldn’t do a terrific job on designing it (he did — as ever), but I wasn’t happy with my own kind of input — and thus, the way we ultimately communicated with you, the reader. The table contained a random selection of stats I considered vital, shifting between per-90-minutes basis and percentages, and so the message must’ve been a bit chaotic. This year, I’ve decided to drop the actual values altogether (I’ll highlight them in text alone if I feel the need), fully committing to percentiles and neat little groupings of co-existing metrics you’ve already grown accustomed to as part of the (largely untouched) pizza charts.
Now, the players will be compared in four areas of their game, so you can quickly see whether they specialize/excel in ball progression, the defensive side of things, and so on. It should be much tidier so you can swiftly take away valuable stuff from the comparison; yet we won’t be shortchanging you on the approach or, dare I say, methodology. That component remains the same.
Also, a slight twist you may notice especially with goalkeepers: even those who’ve started and finished 100% of games they took part in routinely won’t get credited with a nicely round number of appearances. It’s not a mistake, but rather a result of me choosing to look at the actual playing time (including added time) instead of going with the usual 90 minutes = one full game. That’s a speciality of Wyscout, and it’s a welcome one, because while we still (consistently) convert stats to a per-90-minute basis, games simply never have 90 minutes and players make tackles, undergo duels etc. in those extra mins.
Better-conceived metrics: More on the matter in this elaborate introduction.
More educated eye test: For the purpose of tracking goal/chance-creating actions, I watched all 276 Fortuna:Liga games last season. To not go completely nuts and overdose, I’d typically knock off 5-6 over the weekend and then catch up with the rest over the following few days. I can’t say I’ve always paid 100% attention (not by a longshot), but the outcome is me having a much better read on all teams and regular players than I had a year ago.
Indeed, this is about me. I’m not telling you this to brag and suggest I’m better than you; I’m telling you this to suggest I am better than 2021 me — and that I feel I can now trust myself more. When I talk to fellow football pals or colleagues, they often express disbelief I’m doing this to myself. Fortuna:Liga isn’t the best league in the world, after all, and surely even I must see it.
Believe me, I do.
Yet, I didn’t sit down for a single Premier League game — the one league I used to watch almost as religiously — and I barely brought myself to care for the Champions League final. That’s because I value eye test over any data I can dig out — using it merely as a complement, as a useful means to balance out my subjective opinion — and I wouldn’t be doing my Fortuna:Liga coverage justice if I just half-arsed three biggest games a week and be done with it. In fact, I’m finding it increasingly more difficult to even enjoy a random football game without knowing the context, being familiar with the players, etc. I don’t know if it’s just my deformation, but it’s kind of a new territory for me.
One player featuring in two datasets with the entire playing sample: You may recall Oscar Dorley featured as part of the “winger” mini-model last summer despite having his playing time almost evenly split between like four different positions — and datasets. I wasn’t happy about it, and I’m still not at peace with it. This summer, there’s no such extreme case, but I still opted for running one player through two mini-models without splitting his playing time — especially when I couldn’t decide if one is a offensive-minded defensive midfielder or a low-sitting attacking midfielder (or simply wanted to see both results). Such was the case of Marek Havlík, Lukáš Červ or Pavel Bucha. Then there were cases of players who featured heavily in two different roles but I wasn’t able to split their game time between both roles and still meet my arbitrary cut-off (900 mins of actual playing time) for each — or I wasn’t willing to do so to not rob myself of a big sample in favour of two dangerously small ones. Such was the case of Tomáš Ladra, Imad Rondić, Yira Sor (CF/W), Jan Sýkora, Emil Tischler (CAM/W), Adam Hložek or Jiří Klíma (CAM/CF).
What remains is my approach to cases where one could be put in unfair advantage if I included all he did across the season as part of just one dataset — that typically goes for centre backs doing an extensive stretch at fullback, or wingers getting dropped to the backline for a longer period. Centre backs fielded as fullbacks (or defensive midfielders) usually rack up way more offensive output that would instantly put them in comparative pole position at CB — same for fullbacks turned wingers. And it can go the other way with blocking shots and other rudimentary defensive stats, too. So I either discard those minutes at the secondary position altogether (Dominik Kostka, for instance, doesn’t get to flex with his winger output among fullbacks; same with Lukáš Vraštil among centre backs), or split the playing time using Wyscout’s filter and include the said player in both datasets — such was the case of Vlastimil Daníček (CB/CDM), Dávid Hancko or Jan Vondra (CB/FB).
What stays the same
My credentials: They do me no favours. I stopped kicking the ball regularly when I was about 13. I stopped doing any sport regularly when I was 18, as I did my knee while playing floorball at a reasonably high level and never returned due to the timing (I was just starting uni in a different city). I’ve only ever dipped into coaching — and it barely counts as it was at youth floorball level. I’ve written about football nonstop since 2011, yet I’d never consider myself a “journalist” or even well-connected within football. Some footballers do follow me— mostly the young ones who speak decent English — but I don’t think it’s over a dozen. I don’t personally know any agents or top flight coaches and while I’m technically part of a scouting network, it’s only getting started and I have yet to recommend anyone to any club. I can’t do shit in Python or any other programming language, and don’t start about graphics.
My day job, in fact, is miles away from footballing industry and I’m quite happy with that. I spend insane hours on analyzing this bloody sport as it is.
What I can offer then, I suppose, are the resources (old, affordable Wyscout subscription), advanced Google spreadsheet skills (fancy functions and all) to help me extract intriguing stuff from raw, largely unintereresting data, and — well — some willingness to put the aforementioned to use on a weekly basis.
My mission: I started covering Czech football in summer 2019 to provide an alternative view of it. I was — and still am — missing a chunk of nuance as part of its coverage. That’s not to say there are no journalists capable of bringing necessary detail to their reporting, it’s just the broken industry not allowing them to do so; Michal Kvasnica gave me this nuanced opinion on Filip Kaloč’s improvement (that I couldn’t personally see) and his colleague Martin Vait or Jan Dočkal of iDnes — to name but a few — show in their deeply insightful interviews or Twitter analysis they think about the sport on a high level, too.
There’s always someone to play your views off of, but strangely enough, I find that even those with a voice can’t quite find it in their actual job of a journalist. Articles need to be short to have readers, so they require shortcuts like “this guy has numbers” (ie. goals and assists), “this guy is quick/slow” etc. that ultimately provide next to no value and do more harm than good.
Longform read on Czech football is notoriously hard to come by — and that’s where I enter (to arguably over-compensate). I can’t personally bring nuance to everything there is to cover, but I’ve luckily been blessed with many capable helpers along the way. Jakub Lebloch, David Rozlivek, Pepa Javůrek or Vojtěch Mrklas are always there to teach me a thing or two about tactics, Jakub Dobiáš is always there to confront me with a second opinion on analytics stuff derived from the superior 11Hacks database, and I have an entire army of fans who get to see the players in person on a regular basis who have their permanent spot in my team previews, as well. Most data analysts would write fans off for their inherent bias; I rather see advantage in them.
It goes without saying my work would be much less digestible for y’all if it wasn’t for Adam Procházka whose easy-on-the-eye graphic templates made for the crown jewel of the first team preview series already. He’s still on board.
The sample size warning: It should be universally recognized by now that sample size matters and you should never EVER compare a 30-start mainstay with a super-sub worth of a mere 6,5 starts once you put everything together. But apparently, this point needs to be driven home over and over fucking again. Even a difference between 10 and 20 appearances is notable, and I’ll make sure to point out whenever a comparison I make is far from perfect. You’d be surprised how much one start against a lowly Karviná can influence your 10-game dataset as opposed to a 30-game one (where it gets lost).
The context caveat: It also matters who you play for. It’s clear that starting in goal for Slavia, for instance, is a very different job to just about any other goalkeeping gig in the Czech top flight. While Slavia coaches may value sweeping and distributing qualities above much else, a less dominant side will mostly lean on the traditional shot-stopping skills of their custodian. This is why our comparison template will look much different this year — splitting one dataset into 3-5 dimensions of filling one player role to see nuance. A player might have a low overall percentile (considering all 17-18 metrics), but still be an elite contributor in one area of his game that can be harnessed.
So when you’re reading the previews themselves, please don’t linger too long on the overall percentiles. They are, in a way, one of the shortcuts I was complaining about earlier. The individual layers are where true merit lies. Those who come on top should, in theory, have less holes in their game/be more effective in all phases of the game than anyone else in their position (which does actually pass my eye test more often than not). But even when your favourite player doesn’t sit in the comfortable 90+ percentile (meaning he’s done “better” than 90% comparable players), it doesn’t mean he’s useless. It quite possibly means he’s not a two-way winger or a very constructive holding midfielder, but that doesn’t take away anything from the fact they may be bloody dangerous/defensively sound options in one way or another.
Finally, one universal caution: don’t believe anything you see straight away and never take any number at its face value. That’s not how any stats, data of any kind works, and my model — however meticulously constructed (and trust me, I’ve gone back and forth on many metrics) — won’t be any different.
Not now, not ever.
If you count yourself among the fans of my, Adam’s or Jakub’s work, please do consider supporting us all together by donating a small amount of money at BuyMeACoffee page. It would be too kind from you and much appreciated by us!
