First of all, I can’t quite follow @Arthur-Bomber-Harris last post. Sorry, but I don’t understand what you’re trying to say there.
All I can say is, that starting ELO is the same for everyone.
Then the easy answer to @oysteilo first post:
Winner and loser points may differ, when one of them (or both) has less than 11 completed games.
This is because the first couple of games weigh more than later ones. I accomplish that with a “K Factor”, which expresses the amount a game is worth.
That K factor is really high for the first couple of games and then gradually decreases. As @gamerman01 said, the exact values are chosen “arbritrarily”, although of course we thought long and hard about them.
That K factor is essential, because new players all start at 1500, which is almost guaranteed not a perfect rating for them (most newbies are worse than that average, but some might be a lot better too). So we need the system to move new players as fast as possible to where they belong. Normally, a game between two equals awards only 25 (or -25) points. If a new player is actually a 1900 or a 1100 player, those 400 points climb or descend would take a long time without that sensitivity Factor.
Right now, the K factor we settled on is:
As you can see, the first 3 games are worth a little bit more than double the later games.
We can talk about these values, they are not set in stone. But I want to emphasize the importance of bringing new players to their appropriate rating asap. The only alternative would be placement matches - this would mean that new players are not rated at all for their first 5-10 games and opponents would receive / lose just a fraction of the normal worth. I don’t like this option for us.
Now the other question concerning playoff ranking / seeding.
@AndrewAAGamer is correct, originally I planned to use the current year only for playoff seeding. He said he also prefers this and compared it to football, basketball, baseball and hockey.
Now that comparison has one gigantic flaw however: All of these sports are in a league system where every participant has a fixed number of games and the exact same opponents. So it makes sense to start a season with a clean slate.
However, this is not the case with our community. With OOB and PtV, players need to have 3 completed games, with BM4 they need 6. But 3 games are not nearly enough to properly rate a player, especially not if one of those games was an upset (an unexpected loss / win). If we had an entry requirement of 10+ games per year, I’d definitely go for a clean slate every Jan 1!
Our game is more like the other two sports you mentioned (golf or tennis), even though they are still a bit different: It’s impossible to burst onto the scene as a complete nobody and expect to participate in the biggest tournament with only 3 games completed. A first time participant of a Grand-Slam-Tournament has proven himself/herself over many matches beforehand in smaller tournaments. We don’t have that luxury.
I think our sports can probably best be compared to boxing, where everyone chooses their opponents and some have only very few matches per year, while others have some more.
I can change the system to rank playoff seedings only according to results in the current year. Which would mean everyone starts at 1500 (for playoff ranking only). But do we want that?
@oysteilo actually gave a great example!
Oysteilo started the year with OOB rating of 1669
He is 2-1 this year, with both of his wins being almost worthless (against dawgoneit), giving him only +3 each. He lost once against the #1 AndrewAAGamer for -16. Which gives oysteilo a final rating of 1659. Which is more or less the same rating he had for the last 7 years.
ArthurBomberHarris started the year with OOB rating of 1542
He went 5-0 this year, although only one of his opponents was really strong (he defeated #1 AndrewAAGamer!). He gained 110 in the process, which means he significantly improved his rating from 1542 to 1652. This is the highest OOB rating he ever achieved.
They are now almost identical in rating (1652 and 1659), with 32 or 41 total completed games. Which means the rating is very reliable, those two players are very likely extremely similar in strength.
Now it’s a personal decision: Do you think oysteilo should get the higher seed because of his slightly higher rating and the fact he mainainted roughly that rating for 7 years? Even though his rating basically stagnated this year? Then the system should stay as it is.
Or do you think ArthurBomberHarris should get the higher seed because he is on an upward trajectory this year? Remember, Arthur is probably not better than oysteilo (they are most likely equally strong right now) and it is the first time he achieved the same level as oysteilo. But if you think the improvement he showed this year is worth more the system should change to let everyone start at 1500 on Jan1.
Do you think AndrewAAGamer who sits comfortably at #1, with only farmboy being SOMEWHAT close should get the top seeding? He went 7-2 and increased his OOB rating from 1798 to 1830.
Or do you think the 5-0 of Arthur and the 4-0 of Booper this year is more impressive and should give both of them a higher seeding, despite both of them definitely being not as good as Andrew?
My personal preference is the first option, which is currently implemented.
But I can absolutely understand if you value recent results higher than overall strength! This is a system that should be backed by the majority of the community so please: WEIGH IN!
