Dafuge & Data: Stats & Summary Season 1

December 28, 2020

As the sun sets on Scarboroughs first season it’s time for an end of season summary. I don’t really want to do too much of a standard season update. There are plenty of other blogs out there that do it and will do it better. I want to focus on the statistics and what I’ve learnt. So I’ll try to keep the summary brief.

How we started

When we started we only had a few players, and some were on insane wages. We trawled the transfer market to see what youth rejects and journeymen we could tempt to Scarborigh with the promise of chips and slot machines. We were planning on playing hoofball so that made recruitment generally easy. Physical one-dimensional bruisers would do for us.

Transfers

From the slideshow, you can see a summary of the players the Seadogs brought in. Youth mainly but a few brutes (Bright, Richards, Weledji), a little bit of pace (Stanley, Gyamfi, Billam), a promising striker (Varian) and a giant of an experienced journeyman (Kayode).

This all played into my brutality ball hands. Go for long throws, sticking it in the mixer, and hacking ankles.

Fixtures and Table

We started off well, and despite a few blips here and there were actually reasonably consistent. We didn’t really flirt with relegation. Instead, we spent most of the season in the top half, around the playoffs.

Things looked a little bleak at times. Especially when Varian broke his ankle near the end of the season. But ultimately we did enough to get into the play-offs.

Play-offs

We squeaked into the play-offs and were woefully underprepared. Our squad was small, Varian had broken his ankle a few weeks before the end of the season and our form had tailed off as fatigue had set in. Kayode had been doing his best but I wasn’t hopeful.

Farsley were far better than us but even then it was a late blow that knocked us out. I was left with that common cognitive dissonance you get in Football Manager. Disappointed to be knocked out, but pleased that we’d over performed so much to beat our expectations. We weren’t meant to be there anyway.

What about the stats?

So as a brief summary? Massive over performance in comparison to our aims of survival. And we never really looked out of place. We played well, got some good results. Broke a few bones, and put the Seadogs back on the map. But this blog is mainly meant to be about the stats, so in summary what have we learnt this season?

Keeper & Right Back

Before this end of season summary my last post pointed out why Lewis Turner’s time at the club was limited. What we also found that £400pw keeper Allen was also letting in more goals than we would expect.

He played every league game and conceded 57 goals. We were expected to concede 43. A difference of 14 which could be a real issue. He’s partly to blame for undoing the good work of the team in front of him. An easy and urgent area to recruit for next season is in goal. We need a statistically safer pair of hands.

Both Allen and Turner would be leaving, or in the case of Allen waiting to leave because of his contract.

Star Strikers, and Striker Saga

Heroes. Absolute heroes.

Varian managed to bag 25 goals, 22 of them in the league. Not a bad return from 21 xG. Especially seeing as he broke his ankle near the end of the season. He was a cool and collected pressing forward and was well fed by simple passes and flick-ons from his strike partner Kayode.

And Kayode. What a man. If I wasn’t married already I would be planning on an East Coast honeymoon with him. He turned 40 this season. 40. Not only did he score each game either side of his 40th but he got 22 goals from an xG of 19. This 40 year old played in every league game. He didn’t start each one or finish each one, but he played a part in every league game.

I think the game seems to reward experience a bit more now. And having him there as a targetman, really worked. He wasn’t the fastest but he was strong, held the ball up well and passed. And then would drift in to tap the ball home from a return ball or cross.

I had big non-honeymoon related plans for Kayode. I felt that with him and Varian I had a strike partnership that would continue to terrify. But he announced he was going to retire at the end of the season. He gave me 3 days notice, and when I begged him not to retire, he ignored me. That was it. He’s still in the game, but as a fitness coach (figures after a full season in his 40th year). He wouldn’t sign for me. Not yet anyway. Another gap to fill – GK, RB and ST.

I was actually a little devestated by this. The sign of getting invested in a game I guess.

Whislt Kayode and Varian were amazing one potential issue we spotted with xG though was that they were the main sources of goals. The next highest goal tally was 10 from back up striker Stanley, and with xG it was also Stanley with 8. Honourable mentions to wingers Billam and Bright. We needed to diversify a little as evidence by the lack of impact in the playoffs once it all fell on Kayode.

Strength in Defense

I won’t labour the point too much in this summary. I kind of covered it in the RB post. But my defense bar the right-back were doing well. This is where KPI’s come in for me. I’m playing against a lot of longball teams, my defenders just need to defend. It doesn’t have to be pretty. No ball playing defenders here.

So I’ve been judging my defenders on blocks, interceptions and aerial presence. Tackles are a bonus. And Richards and Weledji are a strong centre pairing. Jackson at left-back isn’t too shabby either. The stats suggest that I’ve got a good base to work with.

Expected Points or xPts

We outscored our xG, and we conceded more than our xG against. Good or lucky strikers, and a bad or an unlucky keeper would the simplest explanation here. Another would be as others have suggested is that the xG model is a little generous.

But ignoring that for the moment. If you have xG and xG against for every match you can work out what the expected points from a game with that xG are. From that, you can create a ‘justice table’ that demonstrates where teams should be based on performance rather than the result. I’m not sure exactly how FM21 works it out for the end of season summary but usually, the points are given out as a ratio. Rather than 3, 1 or 0 the points are split based on the ratio of xG for each team in each game.

We can see here that we are not far off our expected points in the summary. And that’s great news. It means based on our performances this wasn’t a lucky season. If we keep up the same relative level of performance then we should expect a similar finish. As I mentioned briefly in the last post, teams regress to the mean. If we regress next season we are still in a healthy position. Technically if we regressed to the mean here we would actually increase the points total as we’ve underperformed slightly on the expected points and final position.

Of course, it’s all a little more complicated than that but in principle, it’s good news. We deserved our finish. A complete overhaul to survive isn’t needed next season. We don’t need a complete overhaul next season to survive. We should be able to make gradual improvements and push on.

A brief word about Efficiency Plots

Sorry, this has been building up for a while. These plots, the ones that summarise attack and defense across the league? Clinical, wasteful, leaky, impenetrable and all the rest? Terrible.

They’re not bugged. They’re just pointless unless you have played enough games to fill it with a reasonable amount of data. It needs at least half a season of fixtures to be of any use, and even then it’s ropey. It’s not predictive, it’s descriptive of what has happened. And if you’ve only played a few games, or if you’ve had an easy run against relegation fodder, it’s not going to accurately represent your actual efficiency.

Don’t get me wrong. I like looking at them. At the end of the season they are a great summary of how you have fared in comparison to others. There are some great blogs out there that were using them before they were even part of Football Manager, that use them well. Especially when used for players than teams. But they’re not useful game to game on a team level. At least not to the extent you’d think they are considering the game pushing you updates to them before every match. We won’t be dwelling on these in our stats posts. But we might be using tableau at some point thanks to FMStag.

Season Summary

I’d give myself an A+ considering we are newly-promoted Seagulls. We’ve solved a problem with our Right Back (well-identified it). We’ve also used to stats to single out the keeper and identify that whilst our general defence is good we need to get more xG in (both to replace Kayode and spread the goals around). A new season means new data.

Additional highlights this year have also included a job offer from league 2 Wigan, an extension to the stadium now we know we are sticking around, and constant takeover rumours. Though hampered by various stats issues but I’m looking forward to season two with the Seadogs. We will be stealing the chips of the bigger sides.

Category: Dafuge and Data, expected goals (xG), FM21, Hoofball, Spreadsheets and Views