Just my $0.02. I was on a path to work up the ladder then work got in the way in 2019…so here I’m am again in 2020. Mostly I like the league scoring and it is a very strong indicator of the system’s validity that the best players are the highest rated players each year. So let’s not muck with the system too much.
The main detractor is that it significantly penalizes top-tier players from playing anyone less. I saw this first hand as I was waiting to play and top-tier players were not willing to damage their scores–I get it.
The suggestion that you play low-vs-high tier games in the non-league forum has some merit, but 2 cons:
a) It’s not a league game so the desire to win is not as great
b) These games take alot of time to play, especially if you want to have 8+ games year and not interfere with real life (work, family, etc.) So I can’t imagine a top-tier player wanting to spend much time on games that don’t count. If you have the time, great go for it. Everyone else doesn’t (possible exception due to COVID-19…boo)
So what’s my suggested modification?
Allow 1 game between players 3+ tiers apart (M-2or3, or E-3) to count each year subject to the following adapted scoring:
a) For the higher ranked person
If they win, the game counts as being played towards the 8 games for playoff, but there is no score, i.e whatever their average was before the game, that’s their average after the game.
If they lose, the game counts but the loss is considered to a player 2 tiers lower than their level (M= loss to tier 1 = 2pts, E= loss to tier 2 = 1 pt).
b) For the lower ranked person.
Standard league scoring in effect.
Rationale:
The main goal here is purely personal–as a possibly upcoming new player, I’d like to see games against better players to improve my game. But also as @AndrewAAGamer said, its better for the league – as in making it more competitive, if everyone gets to learn from the better players. I think with this mod, if a higher player is looking for a new game, they will seriously consider playing lower ranked players.
I set the one game limit between any pair of players to allow more players to interact in the low-vs-high games. Every one may want a piece of the Master, and this way a few more might get the opportunity.
Caveat:
The current scoring algorithm constantly updates ranking based upon the wins/loses of your opponents. This could cause a problem in determining if a game should be a “normal” match or a high-low game once all games are tallied at the end of the year. There is also a possible ambiguous situation that could occur if a low ranked player gets “lucky”. viz. 2 players each play 8 games for the year. One was E tier the year before, the other had no record. The E-tier player plays 4 games against E level goes 2-2 (avg=5.00) and has 4 low tier games (Tier 3) that he wins – no score, but his average stays at 5.00 and he remains a solid E tier player. All good. Now the new player plays 4 games against Tier 3 and wins avg=4.00 and 4 games against tier E players. He a good player but lucky dice help him to 2-2 record. avg = 5.00. So his avg for the year is 4.5, which would also classify as E (barely), but would place him out of the top 8, no playoff, (based on 2019 results). Again, I think most everyone would be good with this.
Here’s the rub. For the algorithm, these players have identical records 2-2 vs E and 4-0 vs Tier 3. How to differentiate them? The solution would be to use the prior year’s rank to determine if you are playing a high-low game or normal scoring. This is also important to make this work because if a M-tier player was really on board with this and gave 4 games to new players, who at the time were Tier 2 or 3, but by the end of the year had worked their way into Tier 1 status. Suddenly, that player will see his avg change as those 4 games would now be (wins?) against 4 Tier 1 players under normal scoring rules (6pts). Might not really be a problem as only 2 people had 6.0+ avg for 2019, but I would not want to be the reason to muck up the higher orders.
Lastly, using last years ranking would simply coding the scoring. Quick check at records to see how to score, rather than iterative examination of everyone’s ranking.
Other constraints
