League General Discussion Thread

[axis-dominion]

yeah adam and i keep swapping seats up there. since i’ve transitioned to ptv i am still figuring it out, and in the process i haven’t been able to progress to 2300 and beyond.

[axis-dominion]

the other factor being that it’s hard to make large gains once you’re up there

[axis-dominion]

and one bad dicing causes a rather large drop…

[axis-dominion]

so if i were to play 666 who’s the top E-ranker and win, i’d gain like 2 or 3 points. if i were to lose, i’d drop 29 or 30.

stakes are like an order of magnitude greater if i were to lose.

fun.

[gamerman01]

It’s kind of a bell curve, right? Yes, you get way up there in rare air, it’s hard to stay up. Did you want me to create tier D, then?

[gamerman01]

@axis-dominion said in Post League Game Results Here:

since i’ve transitioned to ptv i am still figuring it out,

That’s a big part of it, I suspect! You’re not 2200 dominant in PtV yet. You have a heel that wasn’t covered

[Arthur Bomber Harris]

@axis-dominion you are 500 ELO points better than 666. You should be winning 95% of the time against him, a player who has 46 wins and 45 losses, the definition of an average opponent. If you don’t think that you could have that type of odds then you aren’t a 2100+ player.

Top players often have to find meta settings to keep up their ranking, perhaps finding the sweet spot for ELO rank difference and have rules like Low Luck to minimize dicing. Or you could just play games and not care about ELO ranking if you find chasing top-positioning takes away from the enjoyment.

[axis-dominion]

so we’re continuing our game, and he just got 3/4 with his DD hits, that’s following a recent 4/4 DD hits. that’s 7/8 DD hits… what’s that, 1 in 400 odds? his luck is just too unreal

abh
666 is actually a fairly decently good player from the games i remember playing against him, so match would not be a walk in the park, all for almost no elo gain. add to that the ever present risk of dice disasters as happened with peirce, makes it way too risky, assuming i care a lot about my rating. in this case i know from experience playing him (unless he’s changed his approach, as it’s been a while) he’s not one of those that counts on luck to win, he genuinely tries to play the odds and gain advantages through positioning, so i think it’s worth playing him again.

[gamerman01]

This is not an argument to not care about your ELO.

I totally understand we’re competitive gamers and when we put a single number on our success it means a lot.

If it helps, remember that from the impersonal “league’s perspective” the top reasons for a rating, in order of importance to the league, is

#1 To maximize Fun
#2 During regular season, players have a good idea the strength of their prospective opponent. Helps a lot in choosing an opponent.
#3 Decently accurate playoff entry qualifications and seeding
#4 Great way to measure whether and how much you’re improving
#5 Sense of satisfaction, bragging rights, etc

[Stucifer]

I will say if I was playing PtV low luck I can’t imagine Axis would get a bid. I would definitely want a bid to play Allies. I believe with a 0 bid you can be at 100% odds for Sealion on G3 with a naval buy G1 (CV, DD, tp) and 11-transport buy G2.

That’s with UK bringing the Gibraltar + Malta fighter back and building 18 infantry + a fighter for London.

[Arthur Bomber Harris]

@axis-dominion Here is a breakdown of BM4 game results that I pulled from the results spreadsheet. Player 666 is 1 win and 20 losses for players with ELO above 1800. You should be able to beat him 95% of the time if he has not massively improved in recent times. ELO looks to be working in this instance.

Looking at your record (since 2020), you have a weakness against low ranked players and it wasn’t primarily due entirely to bad dice. Against the Simon (ELO 1394), you were distracted by other games and didn’t move additional planes to reinforce the main Russian stack, and he got a bit lucky in the 60/40 Bryansk battle that couldn’t be recovered from. Without that mistake it looked like you would have easily won the game. In the other noteworthy loss against me (ELO 1497), you could have easily won if you played more conservatively instead of offering battles which were dicey.

If you want to maximize your ELO ranking, the 1900+ ranged opponents is probably your sweet spot as the occasional blunders hurt you in the matches against inferior players. Sadly that only gives you 3 opponents who will help you improve or maintain your ranking. The challenges of being one of the best in the world…

[gamerman01]

Boy you sure were right when you estimated 95% off the top of your head! That’s great analysis, and great memory of past games.

[Stucifer]

@gamerman01 Quickly estimating a 95% is the strength of the ELO system, as @Arthur-Bomber-Harris noted. Even ELO should mean 50/50 win rate, being evenly matched players. A 500-point advantage is ~91% win rate.

Re: The new ELO-based ranking system

@MrRoboto said in The new ELO-based ranking system:

Now the Factor F is important: I set it to 500.
This means, that a player with an Elo rating 500 higher than the opponent is 10x as likely to win the game.

Our ELO system would be more accurate if we had a larger pool of players, as one of the issues that has been pointed out is that there is a very small pool of players.

Smaller sample size, less precise data, larger standard deviation in the results.

The top ELO guys can play against only a few others that give them a big chunk ELO for a win, but those players are also the most likely to give them a loss.

However, this is also true in games like chess, DotA and StarCraft, where the ultra-high ELO/MMR players have to play and win lots of games against lower-tier challengers to move up only 100 points in the rankings.

And for the MMR systems, usually the top world players are in the 7k-8k MMR range, so 100 points is only a 1.5% improvement vs a 5% improvement at 2000 ELO.

[Arthur Bomber Harris]

@Stucifer I find it fascinating that we probably only have a decade before a good AI is developed for Axis and Allies, one that can consistently beat the top players. Even an AI that can identify all major weaknesses 3 turns ahead would be a massive tool to take gameplay to a much higher level. Moscow falling a turn earlier is a game changer.

Chess has mostly managed to keep their top ranks honest, but I don’t think we will be able to maintain the League for very much longer. We are counting down the days of top human players.

[gamerman01]

I agree. The one I use seems to now have learning mode on now. I talked A&A 6 months ago and it didn’t know much. Now it does. A lot.

[Stucifer]

Counterpoint:
AI might be able to make the “best moves”, but being a dice game and not chess, it cannot “always win”, unlike a chess supercomputer.

In chess, there is no unknown outcome. Taking a piece will always result in taking the piece. That is not true in Axis & Allies, and never will be.

One of my friends’ dads used to play chess with a guy that was ranked. The dad could only win when the other guy was drunk (which was frequently enough).

Chess is a game where the better player, if better enough, ALWAYS wins. A&A as this entire thread has been discussing, is thankfully not. I play A&A in part because I find games where I can lose to someone “worse” than me at the game and I can win against someone more skilled to be more enjoyable, personally.

[mikawagunichi]

And if the promised AI productivity boom actually materializes we should all have more free time and money to travel and reconstitute the league as an over the board league. Stucifer presciently already has the PTV board printed for us.

[crockett36]

@Stucifer said in Post League Game Results Here:

I will say if I was playing PtV low luck I can’t imagine Axis would get a bid. I would definitely want a bid to play Allies. I believe with a 0 bid you can be at 100% odds for Sealion on G3 with a naval buy G1 (CV, DD, tp) and 11-transport buy G2.

That’s with UK bringing the Gibraltar + Malta fighter back and building 18 infantry + a fighter for London.

This seems like a fairly damning assessment of ptv. Someone might want to think about a version 2.

[axis-dominion]

yep, in ptv SL is much easier, i got it 3 in a row and almost a 4th but stu decided it wasn’t worth it.

[mikawagunichi]

@crockett36 said in Post League Game Results Here:

@Stucifer said in Post League Game Results Here:

I will say if I was playing PtV low luck I can’t imagine Axis would get a bid. I would definitely want a bid to play Allies. I believe with a 0 bid you can be at 100% odds for Sealion on G3 with a naval buy G1 (CV, DD, tp) and 11-transport buy G2.

That’s with UK bringing the Gibraltar + Malta fighter back and building 18 infantry + a fighter for London.

This seems like a fairly damning assessment of ptv. Someone might want to think about a version 2.

More of a damning assessment of LL IMO. PTV is by far the most balanced and dynamic of the the 3 global variants.