Anonymous wrote:Anonymous wrote:Anonymous wrote:YSR does appear to be better but also has serious issues. I see instances were teams are ranked significantly lower than teams they’ve beaten handily recently that are also in lower level leagues. So still some problems with the algorithms used. But, nothing’s perfect I suppose.
In YSR - that will result in the winning team improving their ranking and the losing team will see a decrease in the rank. Every match is factored in so it is more or less a real time adjustment as the teams compete, Figure plus or minus 10 spots but it generally does a good job gathering the stronger teams at the top and weaker teams at the bottom.
It’s not as straightforward as that. If the higher rank team(that lost) is a first team and in a high level league, the team may lose a few decimal points but they will not move below the team that beat them.
Anonymous wrote:Anonymous wrote:YSR does appear to be better but also has serious issues. I see instances were teams are ranked significantly lower than teams they’ve beaten handily recently that are also in lower level leagues. So still some problems with the algorithms used. But, nothing’s perfect I suppose.
In YSR - that will result in the winning team improving their ranking and the losing team will see a decrease in the rank. Every match is factored in so it is more or less a real time adjustment as the teams compete, Figure plus or minus 10 spots but it generally does a good job gathering the stronger teams at the top and weaker teams at the bottom.
Anonymous wrote:YSR does appear to be better but also has serious issues. I see instances were teams are ranked significantly lower than teams they’ve beaten handily recently that are also in lower level leagues. So still some problems with the algorithms used. But, nothing’s perfect I suppose.
Anonymous wrote:Looking at the 04 girls jeff cup bracket - you can see right away that if the showcase has relied more on YSR - they could have improve the brackets,
there was a tennessee game with a low ranking in the top bracket that lost every game by 3 plus goals. (and for some reason they had this team play strong teams in the bracket with high ranking numbers exacerbating the problem)
There was a highly ranked team from charlotte put in the 2nd bracket that won every game - one by 5 goals.
Probably all the teams would have gotten more appropriate competition if the organizers would just consult this system for a good estimate of the strength of these teams.
Not perfect but it works pretty well as long as as the teams have enough history.
Anonymous wrote:YSR does appear to be better but also has serious issues. I see instances were teams are ranked significantly lower than teams they’ve beaten handily recently that are also in lower level leagues. So still some problems with the algorithms used. But, nothing’s perfect I suppose.
Anonymous wrote:Anonymous wrote:Anonymous wrote:Pipeline 04 girls ECNL is ranked low and smashed much much higher teams in jeff cup.
Pipeline is a bit of a special case because they've played hardly any league games this year - presumably becvause of covid - so very difficult for the ranking system to work without data.
However - even despite that - they're ranked #2 in MD with a ranking score of 35.23. They beat TSJFCV 3-1 (ranking 37.02 - less than 2 goal difference), East Meadow 2-1 (ranked 36.04 - less than 1 goal difference), and NC Courage ECNL Academy 2-1 (ranked 35.39 - basically the same ranking). The only result that is even a little bit surprising is the first - but even then is well within the bounds of normal. Soccer games involve a lot of randomness. The best team does not always win. A team really has to be a LOT better to always win. If the difference in ranking is more than 2 goals the better team will usually win - maybe 70%. If it's more than three goals, this goes up to say 80% with 10% ending in a tie.
A ranking system cannot do the impossible and predict which team will win - because
(a) the better team doesn't always win
(b) teams are improving, or deteriorating, or they just get hot for a tournment or even just lucky.
(c) players get injured, coaches change etc. etc.
A ranking system is limited to predicting which team is more likely to win based on its track record to date.
When I look at YSR - if the teams are within a goal then either side has a good chance to win. Anything within two means that the lower ranked team absolutely has a chance to pull off the upset. When the difference is over 4 goals, only then do things start to get into the realm of "just not going to happen". And yes one has to be more suspicious of the ranking for a team with few recent games in its record.
You truly love this stuff man. Lol. You can just look at the national rankings to see who the Top PA teams are...regardless of West and East.
.Anonymous wrote:Anonymous wrote:Pipeline 04 girls ECNL is ranked low and smashed much much higher teams in jeff cup.
Pipeline is a bit of a special case because they've played hardly any league games this year - presumably becvause of covid - so very difficult for the ranking system to work without data.
However - even despite that - they're ranked #2 in MD with a ranking score of 35.23. They beat TSJFCV 3-1 (ranking 37.02 - less than 2 goal difference), East Meadow 2-1 (ranked 36.04 - less than 1 goal difference), and NC Courage ECNL Academy 2-1 (ranked 35.39 - basically the same ranking). The only result that is even a little bit surprising is the first - but even then is well within the bounds of normal. Soccer games involve a lot of randomness. The best team does not always win. A team really has to be a LOT better to always win. If the difference in ranking is more than 2 goals the better team will usually win - maybe 70%. If it's more than three goals, this goes up to say 80% with 10% ending in a tie.
A ranking system cannot do the impossible and predict which team will win - because
(a) the better team doesn't always win
(b) teams are improving, or deteriorating, or they just get hot for a tournment or even just lucky.
(c) players get injured, coaches change etc. etc.
A ranking system is limited to predicting which team is more likely to win based on its track record to date.
When I look at YSR - if the teams are within a goal then either side has a good chance to win. Anything within two means that the lower ranked team absolutely has a chance to pull off the upset. When the difference is over 4 goals, only then do things start to get into the realm of "just not going to happen". And yes one has to be more suspicious of the ranking for a team with few recent games in its record.
Anonymous wrote:Anonymous wrote:Pipeline 04 girls ECNL is ranked low and smashed much much higher teams in jeff cup.
Pipeline is a bit of a special case because they've played hardly any league games this year - presumably becvause of covid - so very difficult for the ranking system to work without data.
However - even despite that - they're ranked #2 in MD with a ranking score of 35.23. They beat TSJFCV 3-1 (ranking 37.02 - less than 2 goal difference), East Meadow 2-1 (ranked 36.04 - less than 1 goal difference), and NC Courage ECNL Academy 2-1 (ranked 35.39 - basically the same ranking). The only result that is even a little bit surprising is the first - but even then is well within the bounds of normal. Soccer games involve a lot of randomness. The best team does not always win. A team really has to be a LOT better to always win. If the difference in ranking is more than 2 goals the better team will usually win - maybe 70%. If it's more than three goals, this goes up to say 80% with 10% ending in a tie.
A ranking system cannot do the impossible and predict which team will win - because
(a) the better team doesn't always win
(b) teams are improving, or deteriorating, or they just get hot for a tournment or even just lucky.
(c) players get injured, coaches change etc. etc.
A ranking system is limited to predicting which team is more likely to win based on its track record to date.
When I look at YSR - if the teams are within a goal then either side has a good chance to win. Anything within two means that the lower ranked team absolutely has a chance to pull off the upset. When the difference is over 4 goals, only then do things start to get into the realm of "just not going to happen". And yes one has to be more suspicious of the ranking for a team with few recent games in its record.
.Anonymous wrote:Pipeline 04 girls ECNL is ranked low and smashed much much higher teams in jeff cup.
Anonymous wrote:Anonymous wrote:And I still stand by my original statement that rankings are goofy. Yes, I do check a rival teams game history at a tournament. Most of us do. It’s useful for that. But when my kids team consistently beats the #1 team in PA, the #3 team in the nation.....so forth and so on....I know it’s BS.
Being the top team in a state doesn't necessarily mean all that much. What was the difference in ranking? I'm very surprised if your kid's team consistently beats the #3 team in the nation ranked by YSR (Gotsoccer is different). A team getting consistently beaten by anyone is not the #3 team in the nation. Even one loss would knock them off that spot.
For example I just picked U14 boys at random. Here are the top 6 teams last losses:
Rank Last loss
1 Nov 19
2 Oct 19
3 Oct 20 (and then Jan 19 before that)
4 Mar 21 (and then Mar 20 before that)
5 Nov 20 (and then Oct 19 before that)
6 Jan 20 (and Sep 19 before that)
None of them have more than one loss in an entire year and three of them haven't lost in over a year.