What I want to see is an analysis of how likely it is that the winner of a match/tournament is also the best team. Basically attaching a p-value to soccer. Then analyze how the rules of the game can be changed such that this p-value is increased.

▲

rkangel 2 days ago | parent | next [-]

I think there is a "best" p-value, and I don't think it's the "highest" p-value.

You don't want it too low, because then quality becomes meaningless. You do want to give good results to good teams. But there is also don't want it to be perfect - you want some unpredictability in sports. You don't want every match to be a foregone conclusion, and you want every supporter to be able to have some reasonable hope.

There is some data suggesting that one of the reasons that English football is popular is because it's low scoring. This increases the chance that random variation gives an "incorrect" result. In this hypothesis, unpredictability adds excitement and builds popularity.

The NFL achieves similar results a different way - various forms of consistency and negative feedback (salary cap, draft order, schedule) to keep teams very close in ability. This means that small differences like a game plan for a particular week can regularly affect results, and keeps predictability low.

	▲	ta1243 2 days ago \| parent [-]
		> you want every supporter to be able to have some reasonable hope. The England game a couple of nights ago. On paper England were better before hand, but Italy were winning until the last minute of injury time. Some England fans had already left the stadium, then in goes the goal, then extra time, then just as it looks like another horrendous penalty shootout there's a foul in the box and England take it, then the goalie stops it, but then it goes in. Certainly a rollercoaster for both sets of fans.

▲

owebmaster 2 days ago | parent | prev | next [-]

What's your intention, the best team should always win? You would ruin football, that's the magic of it.

▲

xnorswap 2 days ago | parent [-]

That unironically is the opinion of many people.

It comes up again and again, and is a culture clash that is not limited to but particularly prevalent between US and European perspectives.

US sports tend to have less meaningful "regular" seasons, which just seed "play-offs", which themselves often have "Best of X series".

All of that is designed to maximise the chance that the "winner" and the "best team" are aligned.

Meanwhile in UK competitions, an entire yearly competition can be decided by a bad 90 minutes, such as ManU losing to York City, something the fans of both sides likely still remember 30 or so years later.

This argument frequently plays out in e-sports, which still try to find a good balance between the two, with the "best players should win" crowd wanting anti-climatic double-elimination, and the "Let's have more meaningful games" crown preferring single elimination.

"Competitions should be designed to find who the best team is" is a statement that many would agree with, but "Competitions should provide excitement and allow for upsets" is one I think is just as important, if not more so.

Another similar culture-clash is the concept of relegation versus franchising, as well as the concept of "drafting" in a (failed?) attempt to even out the competition.

	▲	Certhas 2 days ago \| parent \| next [-]
		I would argue that salary caps + drafts are very successful at evening out the competition. European football leagues have come to be dominated by very, very few clubs. Bayern, Barca and Real, PSG, to a lesser degree Manchester City, have been absolutely dominant in their domestic leagues. In the case of Bayern winning 15 of the last 20, and 12 of the last 13 titles. PSG has won 11/13 most recent titles. In the meantime the NBA has had 7 different champions the last 7 years.
	▲	rkangel 2 days ago \| parent \| prev \| next [-]
		> US sports tend to have less meaningful "regular" seasons, which just seed "play-offs", which themselves often have "Best of X series". Except if you look at the NFL - the most popular sport in the US by far - the playoffs are "Best of 1". The NFL also enforces very close parity which gives a lot of unpredictability. You combine those and you get a lot of upsets.
	▲	philwelch 2 days ago \| parent \| prev \| next [-]
		I think regular seasons are actually better mechanisms for discovering the best team because playoffs always suffer from low sample size. A bad 90 minutes is enough to ruin more than one playoff round in the NFL and probably enough to make the difference in the NBA as well. Playoffs are better optimized for drama.
	▲	Rotoke 2 days ago \| parent \| prev [-]
		[dead]

▲

abdullahkhalids 2 days ago | parent | prev | next [-]

There exist metrics like ELO or GLIKO that can be used to rate the overall performance of teams. This is usually a good predictor for the overall winrate of a team A. But doesn't predict as well whether A will win against a particular B. This is because there is a lot more going on in particular matchups that is difficult to capture mathematically. It is not uncommon to see A beating B consistently, B beating C consistently, and C beating A consistently.

Perhaps a neural network can be trained on a lot more data to make predictions for each match, and I imagine who gamble a lot have probably done things like this already. But then you don't have a number to optimize.

▲

smcl 2 days ago | parent | prev [-]

How are you defining "best" here?

▲

amelius 2 days ago | parent [-]

Best can be defined as the team that wins most often if a large number of matches are played (between the same two teams).

▲

smcl 2 days ago | parent [-]

The winner of a tournament is by that definition the best, it is basically a tautology.

▲

Certhas 2 days ago | parent | next [-]

If I have a coin that shows tails 55% of the time, then there still is a 45% chance that heads wins.

If heads and tails play many games against each other, then the probability for tails to win the overall "tournament" goes to 1. But football is a sport of relatively few games in cup tournaments and low scores (this relatively high variance). This is very conducive to upsets even if we assume perfectly independent probabilities.

Compare to Basketball play offs, with best of 7 Series and on the order of 100 "goals" per game.

That's maybe why in football the league title is more prestigious than the cups, while in basketball the regular season is not considered anywhere near as important as the play offs.

	▲	smcl 2 days ago \| parent [-]
		> That's maybe why in football the league title is more prestigious than the cups Tell that to supporters of Scottish Cup Winners 2024/25 Aberdeen FC (I am a supporter of Scottish Cup Winners 2024/25 Aberdeen FC) :D

▲

amelius 2 days ago | parent | prev [-]

I mean a large number of matches played between the same two teams.

You can compare it to how it is done in medicine. Imagine a match between a drug and migraine. Would we only do a single test to determine if the drug "wins" against migraine? Of course not. We do many tests and determine a p-value. We can do the same thing in soccer.

Now, of course we cannot do this in a real tournament (it would take too long), but we can draw conclusions from such a test, or several such tests.

	▲	smcl 2 days ago \| parent [-]
		The problem there is that there is often far too much variation in the teams year on year - squads change as players get bought, sold or retire, managers come and go. And when you do find a pattern I can virtually guarantee that it won't be due to some novel gamesmanship, but rather finance. It is no secret that the most successful teams are those who are wealthy enough to able to buy (and pay) the most sought-after players and coaching staff, and build and maintain the most sophisticated training facilities (and a whole host of other smaller things that cost money). If we take a look at the finalists of the Europa League this year you can see that the money they both spent on their squad dwarfs all of their previous opponents in the competition combined: https://www.reddit.com/r/soccer/comments/1krtgtt/combined_pu... (ideally you'd factor in wages too but that's harder data to get hold of, but it'll paint a similar picture of not a more polarised one). What is really interesting and far more worth studying is that there are some very fun outliers - clubs like Bodø/Glimt who have a miniscule annual budget but have overperformed in European competition in recent years. They reached the semi-finals of the Europa League - beating Lazio, Olympiacos, Porto along the way - and in previous years have had similarly deep runs where they battered Roma 6-1 and beat Besiktas and Celtic both home and away. All of the teams I mentioned Bodo/Glimr defeating will have an annual budget that is 10 time theirs (or more) and will frequently make high profile international signings, while Bodo spend frugally and have a predominantly (if not all) domestic squad. I mentioned Manchester United and Tottenham - they're also outliers worth studying, but in another way. They were both utterly woeful in the league last season despite having astronomical budgets. If you can crack what causes a Bodø/Glimt, a Manchester United or a Tottenham then you'd be a very valuable addition to the backroom staff of any football club with a desire to punch above their weight...