European Coaster Poll 2024 (last 16 open)

[Martin]

The penultimate round of the league phase has begun

TS Poll Gameweek 7

forms.gle

 

[Martin]

It's the final week of the league phase. It's still incredibly tight so every vote matters!

TS Poll 2025 Gameweek 8

forms.gle

 

[Martin]

The winter closed season is almost over and half term events are just around the corner, so it's time for things to get serious. After 8 rounds of matches, the league stage has come to an end and the top 16 rides enter the knockouts!

Here's the league table, sorted by points, Buchholz coefficient and goal difference. As you can see, there's a couple of shocks, including previously perenial winner Helix not even making it through the league phase!

GD Buchholz Points
1 Voltron 121 118 24
2 Gotham City Escape 48 121 21
3 Zadra 90 119 18
4 Wodan 23 105 16
5 Lech Coaster 45 100 16
6 Taron 30 129 15
7 Taiga -16 125 15
8 Ride To Happiness 18 116 15
9 Untamed 29 99 15
10 Karnan 25 84 15
11 Toutatis 35 112 13
12 Hyperion 26 96 13
13 Kondaa 30 89 13
14 Shambhala 24 112 12
15 iSpeed -44 90 12
16 Expedition GeForce 26 88 12​

------------------------------------------------

17 Hyperia -8 85 12
18 Black Mamba -23 84 12
19 Goliath -18 73 12
20 Wildfire -13 105 11
21 Helix 4 119 10
22 Katun -37 83 10
23 FLY -27 102 9
24 Oz'Iris -28 99 9
25 Colossus -36 76 9
26 Balder -14 70 9
27 Blue Fire -29 65 9
28 Silver Star -9 85 7
29 Monster -70 76 6
30 Superman -87 75 3
31 Dragon Khan -72 68 3
32 Red Force -43 64 3​

Here's the Round of 16, as usual the poll closes at Sunday midnight and the last 8 will begin next Monday!

 

[John]

Excellent work so far, thanks for organising. A few surprising results in there but without knowing where each ride picked up wins and losses it's hard to say how much of that is just luck of the draw. Helix and Wildfire in particular look to be very unlucky as had they won one of their drawn matches either would presumably have gone though? Similarly a couple of votes either way could have seen Hyperia or Mamba qualify in place of EGF without even needing to change the result of a single "match".

I think "goal difference" is fairly self-explanatory given the format of the competition but I've never encountered Buchholz before. Looks like a way of measuring how well each ride did normalised by the strength of the opposition but I can't quite figure out how it's calculated - wiki suggests two methods, neither of which seem to lend themselves to negative scores, can you give a bit of background as to how it's been applied here?

 

[Poisson]

Hyperia outranking Katun should be illegal

 

[Martin]

Excellent work so far, thanks for organising. A few surprising results in there but without knowing where each ride picked up wins and losses it's hard to say how much of that is just luck of the draw. Helix and Wildfire in particular look to be very unlucky as had they won one of their drawn matches either would presumably have gone though? Similarly a couple of votes either way could have seen Hyperia or Mamba qualify in place of EGF without even needing to change the result of a single "match".

I think "goal difference" is fairly self-explanatory given the format of the competition but I've never encountered Buchholz before. Looks like a way of measuring how well each ride did normalised by the strength of the opposition but I can't quite figure out how it's calculated - wiki suggests two methods, neither of which seem to lend themselves to negative scores, can you give a bit of background as to how it's been applied here?

Here is a pdf with all the league results

Buchholz is indeed a way of measuring relative performance but I think you must be looking at the GD column instead of Buchholz, there's definitely no negative Buchholz scores.

As a bit of an explainer for anyone who's not sure how this all works, the idea of a Swiss system is to maximize how meaningful each round is by matching up similarly ranked entrants. With 32 entrants there simply isn't time for every coaster to play against every other coaster, and a straight up knockout would require seeding and still be unsatisfying with half the entrants going out each time without any real context for their performance or popularity.

As football supporters may know, the Premier League table is already around 90% taken shape by the 10th round of 38, as by then despite the 'random fixture generator' teams usually have played a range of opponents from best to worst. By the end of the season, mid table teams are often accused of being 'on holiday' due to the meaninglessness of their matches against other teams with little to play for, which have little to no effect on the overall standings. The players aren't fussed, the fans aren't fussed, the league table isn't fussed. Late season matches between teams at the top or bottom are often refered to as being 'six pointers' because of how important they can be.

Therefore, if we only have a limited time to run a competition, want every match to be meaninful and for the table to be representative of the entrants actual relative qualities, we should aim not to have a random and predetermined fixture list, but to artificially curate each round of fixtures to have entrants close to each other in the table. The first round was randomised, and then the software chose the rest based on the standings. Each ride would be drawn to the closest performing ride, provided that they had not already matched up. This means that a strong coaster that started off against another strong coaster and lost had a good chance of getting a really easy next round, and so on, until things start to stabilise - look at the state of round 2.

With 32 entrants, around 5 or 6 rounds is where things start taking shape, but even with 8 rounds, as you can see, it's still very tight and could have changed a fair bit with another round or two. Statistically, 8 rounds is enough for the table to be fairly reflective, but it's likely that some entrants will have still had a worse run than others, which is where the Buchholz score comes in. Buchholz adds up the points of every opponent. By example, Helix's Buchholz is quite high for where it placed, suggesting it got stiffed by the draw, whereas Karnan's is really low for a high placed ride, suggesting it got lucky. If the rounds carried on, this score would become less important and eventually, goal difference would take over as the most meaningful way to tie break. As you can see, some of the goal differences are wild, so we're probably far from that point, but I'm no statitician so this is just a guess.