So this just feeds the boot-ball winning at all costs over development disease |
It’s not just score difference. It matters against whom you are playing. Strength of schedule. See Elo Rankings |
sure or also - how about all those clubs that sandbag in tournaments just to appease parents. Now you can see - no **** you beat a team by 4 that you claim as the greatest win ever at the club. |
Clubs can put their teams in lower brackets for wins but all you're doing is forcing your team to win by 5 (just an example) goals to maintain their current ranking. Here's why. If |
Here's why, If you play worse teams the ranking app will expect you to win by a large number of goals. If you win rhe game but not by enough goals its less than the expected outcome and the team goes down in ranking. Conversely if the team that lost had more goals than expected they'll go up in ranking. |
Then don't play down in lower brackets? That's a feature not a bug |
This is something else I really like about the ranking app. Before it was availabe all we had was paid for rankings and tournament/event results. Now because of the ranking app theres an alternative way to view teams. Clubs, leagues, tournaments, etc can manipulate all they want but because the ranking app uses an average of over one year its hard to cheat. (Unless you cheat consistently) |
I agree. Play down all you want but its only hurting your teams ranking if you're not beating then by the expected or more than expected number of goals. |
The biggest flaw i have notice in that ranking app is i have noticed in the younger ages my kids team could beat a top 20 team 1-0 and you see no real difference in ranking then the next week they beat a team ranked 500 12-0 and that win bumps your rankings way up. The first win was a much better win but the app gives way more pts for the second win that wasnt very impressive. The rankings seem to be more accurate the older you get and less of the one sided games exist. |
What youre describing is exactly how it should function. Likely the first game the expected result was 3-1 for the top 20 team. Since you won 1-0 nothing will happen with your team. The top 20 team likely dropped 20 spots and is now top 40. The second game you won 12-0 and the expectation was probably 8-1. Which means your team exceeded expectations by 4 goals which pushed up your ranking and the other team was one goal lower then expectations so their ranking went down. One of the ranking apps issues while not their fault is that some leagues and tournamants cap scores at 6+ goal differentials to limit the impression of blowouts. This holds back teams from gaining in ranking when they score 12-0 wins because they'd be recorded as 6-0. |
It was built by a data scientist using machine learning. It’s very good and better than any other biased ranking system out there (I see you got soccer, top drawer soccer and all the other subjective ranking groups out there with an agenda)….
It is very comprehensive. AI is probably making it even better. Perfect - no. But very good. |
Stop saying AI - no need for AI - it's a straight ELO algorithm that has been around forever. Kids these days!!!!! |
This is correct theres no AI going on. Although if you had access to the raw data you could use it to pull info out. You could also sell coach performance metrics to club owners. Or club performance metrics to leagues. ![]() |
The secret sauce is the dev is a maniac about curating the data. They probably could sell it off, but seriously if you know nothing about this type of thing - it's SUBSTANTIALLY better than any subjective measure before. Yes you can skew it, but over time it's been extremely accurate. |
Hes using webscrapers to pull game result data into excel or a database. Then he massages the data to get into a usable format. Finally publishing the daily results which gets shown in the ranking app. I agree not much work. The hardest part would be keeping track of all the tournamants and random events that are going on. But because they provide links to schedules before the tournament/event it wouldn't be that difficult to add new all the time. Also I'm sure theres less than 20 different schedule formats so once youve got one pulling in it will work for that type going forward. For leagues + tournamants. If you want to throw some roadblocks at the rankinggs app guy. Present schedule data as a json. Or maybe require logins + track who is looking at all the data in a scripted way then block that user. ![]() |