TSL and Basing Cup Announcement

[MHz]

The rampant aliasing is getting really tired now. Obviously it's only mediocrity that can get away with aliasing, since true talent is always identifiable. Nevertheless is confers an advantage in that opponent players have to spend time figuring out how good you are during a game. Furthermore, some aliased guy "shamanzi" pretty much sabotaged a recent round I played by teamkilling me 3x carelessly, and then proceeding to go out 2-10... That's the kind of performance that would normally make me mentally consign you to the irredeemable trash heap of total fuckwad newbies, yet he avoids this censure and reputational damage by being aliased. Guillotine for aliasers imo.

[ogron]

can we just, up front, eliminate aliases from TSL? or at least give everyone the opportunity to register with the bot... and then have no chance to reregister a new name. you pick a name and you play on it the entire time. you may not choose a new alias to register.

is there any reason not to do this other than some asshole wanting to alias for no reason other than to be annoying?

agreed, hopefully they clean aliases up.

[ogron]

IIRC aliasing was always banned. you had to play on 1 name and that's it.

[Zeebu]

can we just, up front, eliminate aliases from TSL? or at least give everyone the opportunity to register with the bot... and then have no chance to reregister a new name. you pick a name and you play on it the entire time. you may not choose a new alias to register.

is there any reason not to do this other than some asshole wanting to alias for no reason other than to be annoying?

[Zeebu]

can anyone bring back TWLM or something similar to tack onto the TSL stuff?

[ogron]

stats that in my opinion should have some extra weight would be:

- player eliminations giving bonus points relative to how close the match is at that moment. these kills are worth far more than your regular kill as they tend to swing the match in favor of one team. first elimination should have more value than the rest with it slowly scaling down. there are exceptions such as 1vs1 or when someone turns around a disadvantageous 1vs2 situation. additional points for being on the winning team.

I had the same thought, I asked BIET about this and he said from a bot perspective it would be very difficult to code and probably unlikely. I agree that if you could factor score and 3v3 or 5v1 would be amazing... a 5v1 KO at 48-34 is worthless, but a 3v3 one at 40-40 is high value.

- early eliminations from a game should be slightly penalized where as staying alive should be rewarded. this should be relative to current game score as dying out at around 30-30 score can be very detrimental to your chances of winning. I know this could lead into some players playing far too safe so only minimal value gain would be preferred.

this one has less value, and I can think of a particular low value warbird who would get max points from this staying alive on 3 every game while his team is on 9.

- I don't agree that grinding 100 games should give someone massive advantage to qualify for playoffs compared to someone playing only 20 games.

quick example would be that someone plays 74 games in TSL. we would look at the average value of that player across 47 games (base 20 + 27 as the player played 54 extra games). this still gives an advantage to players who want to grind out games, but it's not nearly as massive. it also rewards players for being consistent as players with huge performance swings should not be rewarded which is what the current system essentially does when someone plays far too many games. you could tweak these values to be more strict which would be my recommendation.

as far as players going inactive you could start giving a small value penalty to players who don't play for a week with it slowly ramping up to more and more loss. many games with ranking systems use a similar system for inactive players to promote activity.

Not a bad idea, I just worry about adding coding, formula, and bot complexity for small value gains, for a league meant to fill the gap between November and Christmas that starts in 2 weeks.

[MHz]

Agree with Turban re: early KO's (rewarding early KOers and penalising the early KOed.)

Two quick points re: basing ratings that I've borrowed from the SVS statbox.

1) For sharks - an interesting metric would be "wasted reps". This measures both missed spawned reps, and missed greened reps. This reflects more of the "missed opportunity" factor that I think often separates elite sharks from good sharks. It has one draw back in that, in certain circumstances, not repping (and dying with your reps) is the best possible strategy for that death (when terrier is behind enemy mines.) But generally I feel this would be a slight improvement on RPD (where mistakes and inefficiency can be flattered by greened reps.)

2) For spiders - an interesting metric might be "damage inflicted". This would give spiders who make selfless but impactful plays (draining sharks to waste reps, for example) a small edge over spiders who tend to cherry pick easy kills.

In a similar vein, you might be able to produce a ratio between "damage sustained" and "deaths" to give a kind of resilience factor. It seems to me that the art to good spidering is being able to sustain high damage and not be killed - and being able to straddle that fine line for maximum efficiency.

[LeeRomeno]

Good points from turban. We tried similar initiatives in EG, 2 things which really ruined the experience:
1) aliases
2) annoying fcks who were added to teams, but who in reality, in vp system should have negative ratings, as they fcked up games on purpose. Lets say going on suicide run and eliminating yourself 0-10 within first 5 minutes, then aliasing and doing the same shit next game.

Hopefully tw has less of those idiots (although i have seen some transfer here from EG aswell) and some things are done to prevent that.

Other than that, never really played this type of leagues before, so waiting for it.

[Turban]

stats that in my opinion should have some extra weight would be:

- player eliminations giving bonus points relative to how close the match is at that moment. these kills are worth far more than your regular kill as they tend to swing the match in favor of one team. first elimination should have more value than the rest with it slowly scaling down. there are exceptions such as 1vs1 or when someone turns around a disadvantageous 1vs2 situation. additional points for being on the winning team.

- early eliminations from a game should be slightly penalized where as staying alive should be rewarded. this should be relative to current game score as dying out at around 30-30 score can be very detrimental to your chances of winning. I know this could lead into some players playing far too safe so only minimal value gain would be preferred.

- winning games is obviously worth a decent chunk of points. this is a no brainer. current simple formula might be a bit too heavily skewed towards it.

- I don't agree that grinding 100 games should give someone massive advantage to qualify for playoffs compared to someone playing only 20 games. it makes playoffs feel cheap and thus in my opinion the formula needs to be reworked. one alternative could be ranking players, at least for playoff purposes, based on their average value across X games with that number going up by one for every two games played after reaching required minimum amount of games played. we could require players to play, as an example, 20 games in order to be eligible to qualify should you anticipate a lot of activity from the community.

quick example would be that someone plays 74 games in TSL. we would look at the average value of that player across 47 games (base 20 + 27 as the player played 54 extra games). this still gives an advantage to players who want to grind out games, but it's not nearly as massive. it also rewards players for being consistent as players with huge performance swings should not be rewarded which is what the current system essentially does when someone plays far too many games. you could tweak these values to be more strict which would be my recommendation.

as far as players going inactive you could start giving a small value penalty to players who don't play for a week with it slowly ramping up to more and more loss. many games with ranking systems use a similar system for inactive players to promote activity.

just a quick post before work.

[Poseidon]

to give some fresh perspective:

imagine the bot for TWDD could record how many shots a player had fired against them:

- one player finishes 5-7 and had 200 shots fired against them
- another player finished 13-10, went out 2nd or 3rd, and had 45 shots fired against them

who gets more value?

[MHz]

Correllation between personal score and utility to your team/likelihood of winning is pretty strong. Good enough for this game anyways.

However, +7 seems a bit too generous for a win.

I feel like going 16-10 on the losing team is better than going 10-10 on the winning team. To give some perspective, Ease averaged 11-9 in TWDTJ main season. The idea that you should be given the equivalent of an additional 7 kills per round for the win seems too much to me.

I also think 25 point average is a little steep for medal qualification. Taking the example of Ease again, he would've averaged roughly 17 points per round this TWDTJ main season.

[Poseidon]

stat-based league HAHA

anyways sign me up for basing cup but not to captain, cheers

[MousE]

15% show rate for this event

I am personally holding you accountable to that!

[ogron]

saiyan i have zero attachment to the format of TSL, if you and Turban have a way to improve it and make it about winning, i'm all for it.

as a competitive person, i am all for ELO, only counting winning, and ignoring stats, as i think stats are mostly worthless. i post stats for entertainment purposes. winning is the only real tool we have, short of having SportVU cameras in every arena rating every play you make against a database and determining its impact on the game. TSL uses stats because they're available and very good at parsing talent on small sample sizes.

but you need to overcome some issues.

1) there's nothing worse than a great formula that uses data with small sample sizes that make it useless.

2) i think they're looking for a league that can be finished by Christmas? i mean you can discuss it with Racka, he's in charge of TSL, maybe you can extend it to get the sample sizes you need, but personally my volunteering for this stuff is for 2021 only. If Racka wants to run it into 2022 so everyone gets 300 games, ok.

3) Ease doesn't like "only your best 20 games count" either, and says it would be more accurate if it was your averages or counted every game. Again, I don't disagree, but this opens players up to trolling, it increases toxicity because now every game counts and anyone doing anything counter to your success is permanently impacting your rating. If it's just your 20 best games, and someone throws or trolls, you can just move on. It also keeps players who are at the top of the leaderboard playing instead of "sitting" on their rating.

4) What do you do about sitting on your rating? We want to promote activity. You see this in other games and elim where people get the rating they want, then are highly incentivized to not play the game. We want people playing. Other players still trying to rank need guys to step in and play, if players are sitting on their rating it gradually decreases activity.

5) Are there enough people willing to play 300 games to truly make this worthwhile? What happens to players 100 games in who realize they have no shot, will they keep playing or just quit? If the league becomes mostly good players who can hit the Top 10, does that make having played most of your games early in the season against worse competition better than late in the season where others are better?

6) Is there enough granularity to separate players from each other meaningfully with the formula and amount of games played?


The current format features design choices that were meant to solve a lot of these problems, but I'm happy to play a nerdier and more win/elo-centric league if you can make it work.

[Turban]

saiyan has some fair points that can definitely be fixed prior to the season starting. The purpose of this TSL is to keep activity healthy during the holiday season for the next month (Nov 14 - Dec 14). Running TWDT back to back seemed a little redundant and I feel like this is the perfect time to bring back TSL as a happy medium.

Turban I understand the hesitancy but the community can make the league as good or as bad as they want it to be (IE: check season 1-5 to see the drop off). The finals have potential to be as competitive as we want it to be, think about the possible line ups.

TSL checks all the boxes for where the game is at currently. EVERYBODY gets to play and the BEST 10 warbirds and javs get to compete for the titles. The season is 1 month long in total and can be a huge success if the community buys into the league.

Another big downside for me is that I have essentially no free time during weekdays (and even some weekends are fully booked with work) and games tend to run, at least historically speaking, at quite late hours. It is simply not realistic for me to reach 20 games played to even qualify for the playoffs which is a huge drawback. I have skipped pretty much every season in the past with the exception of the first one where I played a few games and subsequently decided that the league isn't for me.

In the end I am sure TSL will run fine and there will be enough players interested in it. I simply won't be participating for various different reasons.