(10,622+627)/627 is roughly 18. So a boys mls next team would need to play 17 times other mls next teams before they get to play an ecnl team.
I would like to know why they think that’s enough inter league play for the algorithm to work.
If you really want to know things, you have to at least try and understand them when people explain them to you. Stubbornly continuing to believe falsehoods based on a lack of knowledge is a trait we all try to have our kids grow out of. If you have actual questions - ask them!
The required amount of interplay to calibrate effectively between leagues is much, much smaller than you believe is necessary. It's like Vitamin C. Have none for many months? Get scurvy and even die. But have as little as 10 mg every day or so, and you'll never have a problem. And a bit of it is supposed to be good for colds. So people take 250 mg. Or 500 mg. Some even take 1000 mg or 2000 mg thinking if a little is good, more must be better. But over a pretty low threshold, any more doesn't actually do anything (other than support supplement companies). People are just going to pee out any extra anyway.
Interplay is the same way. Enough interplay doesn't have to happen so it's as if it really is one pile of homogenous teams, all playing each other a roughly equal amount of times. A team gets it rating affected every time it plays another ranked team. It is affected by all of the opponents of its opponents. Even within a very small ("closed") group, they all are stratified by their results in a way that best represents the relative performance of the team. And once any of their opponents, or their opponents opponents, start to have game results from other populations - those results all have an effect on adjusting the ratings. In cases where the actual results are way off - that result moves ratings of all teams in the league much more quickly. In cases where the actual results are pretty close to expected - that result ends up not needing to adjust ratings much at all. It doesn't have to be any of the top teams playing outside. It doesn't have to be any of the bottom teams playing outside. Any/all of them will have the appropriate effect on the group and each other.
Some regional leagues are stacked with all good teams. Looking at a league in its entirety is like an average. There will be outliers. The outliers will be most noticeable on top clubs that only play other top clubs.
All this quickly will change when teams play against teams they dont normally play in Tournamants, Showcases, or Finals. Its also a good way to see that a feedback loop exists.
I just don't see how this is as impactful as you believe it to be, at least not for the reasons you're attributing. By the time it gets to playoffs/finals in these top leagues, there certainly is a dip in predictivity compared to regular season. It's now going from a full group of teams with skill A to skill Z, and cutting them down to a subgroup of teams with skill A to skill F. The relative differences between them are smaller - making the results that much harder to predict. It happens reliably, every season.
If one really wants to find a set of teams that all have a rating that is well off "reality", defined as the expected rating nationally for a team that is objectively performing at that level, they need to look at the very youngest teams when they are first established in brand new leagues. A town has a brand new set of teams in travel ball, say 8 of them, all U8. All of them have zero history, and all of them are starting to play nobody but each other. None of the teams have a rating at all. They all are given a hidden starting rating tied to their age. They start to play each other, and teams start to stratify themselves in the standings bracket. Once they've played 5 or 6 games, the ability to calculate expected performance becomes statistically significant, and they start to pop in to the Ranked teams rather than Unranked. If at this point, there still have been 48 games played, with zero extra tournaments or anything else in any of their histories, then the teams themselves are the only ones that can directly affect their rating.
But even so - if you take the average score for that entire league, and compare it to the average score for the entire league of the town next door and their U8 teams - it's likely to still be awfully close, as 8 year old kids here as a group are pretty similar to 8 year old kids over there as a group. Playing in the small closed league may be seen as a feedback loop by some definitions - but it's not a feedback loop that continues to emphasize incorrect or unwanted information (like a feedback loop in a mic/speaker that results in those high frequency squeals).
For this same thing to happen in a regional league at the older level, when more people might agree that ratings/rankings could be perceived as more valuable, it would have to be made up of all entirely new teams - with zero history among all of them. All of the teams would have to have been started with a generic rating, and none of them would have played any games outside of that brand new league. I'm sure that it happens from time to time, perhaps at the U13 age when new teams are established, but it's a fairly limited time span before the collective game history across all teams gets wide enough to bring things in line as needed. And once it does - it's not like that calibration goes away in a short time and needs to be "rebuilt". The continued results of the team continue to adjust the rating each game, and they survive season over season rather than having to start from scratch every 6 months.
