S19-S21 Advanced TWLD Stats

[ogron]

Here are some more advanced statistics. These stats only count games between the Top 4 squads in TWLD during Season 19, 20, and 21 (as of Week 5). This means only games between these squadrons were calculated:

S19: Fierce, Thunder, Cobra, Sweet
S20: Fierce, Burn, Cobra, Potenza
S21: Dice, Thunder, Cobra, Force

The goal of these stats is to exclude stat-padding against lower-tier squads, and see how well players perform against the best squads with good teammates around them.

Stats never tell the whole story, and there are certain intangibles like saving a low energy teammate from a rush or luring that cannot be accounted for, but I think these are a better look at the stats than simply using the raw stats from all games.

Legend:

GP -- Games Played

Conf. -- Confidence in your stats. I set this at 10 GP, as this was the cumulative average. If you've played 7 games, your rating is multiplied by 70%, reducing your score. Anything above 10 games played and you get your full rating.

KPG -- Kills Per Game. These should be pretty accurate as I manually removed most late-game sub-ins.

DPG -- Deaths Per Game

Win % -- Percentage of games won

Alive % -- Percentage of the time the player stayed alive when his team won. This stat is relevant and factored into the rating.

Plus-Minus -- Kills minus Deaths

Rating -- (((KPG*10)+(Alive%*40)+((Plus-Minus/GP)*20))*Conf.)/1.999

[Ricko]

well I suck....

...Although looking at win % I'm pretty much 20%+ lower down than everyone above me, so I'll do the gracious thing and blame every teammate I've had in the last 3 seasons

[ogron]

well I suck....

...Although looking at win % I'm pretty much 20%+ lower down than everyone above me, so I'll do the gracious thing and blame every teammate I've had in the last 3 seasons

I don't know that I would qualify having the 8th best stats among active players as "sucking", especially given that I think win % is more indicative of the quality of your team than the quality of the player, which is why it isn't factored into the rating. TWLD is the least interdependent of the three leagues, where you can have good stats on a bad team (Racketeer is a blistering 78-59 on Paladen this season, and Jones had an amazing 111-79 record on Vikings last year), so while in basing I prefer metrics like adjusted plus-minus using win differential in minutes over raw K stats, in warbird I think using those is fine.

Also, context is important, and playing on a team that lost 68% of its hardest matches is probably a more difficult environment to excel in than playing on a team that wins over 50% of its matches, as you can get a few easy kills at the end of wins that aren't available when your team is losing. Cobra also played Fierce 9 times during Season 19, and you probably played every one of them, whereas someone else might have played them once and their other matches, while nevertheless against good opponents... it's not the same thing as playing those tryhards on vent.

And you have an excellent record versus Fierce, going 85-83 versus them (13-10, 8-8, 10-9, 13-9, 7-10, 4-7, 12-10, 5-10, 13-10), whereas even players ranked higher than you haven't held up well against them. For example, Iron Survivor is 35-44 against them over the past three seasons. There's not a lot of people who have positive records versus truly elite competition.

The eye test is still the best methodology, there are players on this list I would rank higher based on how they play (luring, covering, managing deaths, teamwork, trading out of bad situations, not crossing, playoff clutchness, etc.) but I think stuff like this at least provides a sort of reference point. You can argue x player ranked y should be z spots higher or lower, but there's an upper and lower bound that stats provide to ground that conversation in facts.

[Crescent Seal]

did u include stats for Madara Uchiha 4 me

[Attacks]

Good read thanks for this

[ogron]

did u include stats for Madara Uchiha 4 me

I did, yeah. I did a REPLACE ALL sweep and changed all the aliases I recognized. If you want a better look at my accounting, here are the games of yours I included in the sheet:



Nice S21 btw, elite 10 straight rounds either going + or even 'til Cobra -- Cobramin fked us all this week

[Crescent Seal]

struggle has been real. cool stats tho.

[ogron]

thx.

and good luck w/ the streak. u r the current twld leader of staying alive in every win:

22gp -- Cres (100% alive)
19gp -- Welt (100% alive)
17gp -- Riverside (100% alive)
15 gp -- Racka (100% alive)
15gp -- Iron Survivor (100% alive)

the streak lives.

[Crescent Seal]

streak lives

[Ricko]

appreciate this claus, good to know I'm 85-83 vs Fierce overall!

Your stats made me look over my TWLD stats over the last 6 seasons and it's pretty cool (by cool I mean sad as fk cool), especially when I look at my total rec over the last 3 seasons and minus the stats you've put in here vs squads from the top 4....vs every other squad I've gone 263 kills 123 deaths....so I'm good at stat padding.

Wish we could see every TWLD competitors recs vs each squad like you just did for me vs fierce,that'd be a cool feature - i.e. I could see who has the best rec vs cobra all time etc. But that's alot of work and unlikely to happen any time soon.

[ogron]

appreciate this claus, good to know I'm 85-83 vs Fierce overall!

Your stats made me look over my TWLD stats over the last 6 seasons and it's pretty cool (by cool I mean sad as fk cool), especially when I look at my total rec over the last 3 seasons and minus the stats you've put in here vs squads from the top 4....vs every other squad I've gone 263 kills 123 deaths....so I'm good at stat padding.

Wish we could see every TWLD competitors recs vs each squad like you just did for me vs fierce,that'd be a cool feature - i.e. I could see who has the best rec vs cobra all time etc. But that's alot of work and unlikely to happen any time soon.

Thanks!

I mean, it's not that stomping lower tier squads is worthless, going 19-4 versus Gantz or whoever shows carry potential, it's just that it typically doesn't translate well to playing more elite, intelligent, and disciplined opponents.

Yeah i would love to see every TWLD competitors recs vs each squad, too, but it's probably way too much work. Turban and I do love a challenge, though, and we've used TWL stats as an excuse to learn excel over the years, so if people want more TWLD stats, i'll see what i can do.

TWLD has always been my favorite league to cover (see Riverside, Racka, and Ricko: Race for the Scoring Title), since I didn't play it and could cover it with an objective reporter's eye from afar with no real stake in it (whereas any basing stats I produced seem to be eyed with suspicion by my forum stalkers since I play TWLB), but had the advantage of having been a WB before and knowing what to value. That has kind of been lost a bit sadly due to depth issues S20 and S21 forcing me to play sometimes, even though I'm just a mediocre support player far back on the depth chart.

Season 19 will always remain one of my favorite TWLD seasons to follow, since I played 0 games, and just got to enjoy a nice resurgence in warbird competitiveness (Season 18 sucked), and that massive wb/jav schedule provided a ton of games to follow and a great statistical sample.

A reasonable future project would be doing what I did here for Season 16, 17, and 18, and showing them separately + combined. The issue is that S19-S21 featured 4 clearly top-tier squads -- it wasn't hard to choose the cutoff at all. I have a strong suspicion it's not going to be that simple going further back.

[ogron]

S16-S18 Advanced TWLD Stats

So, S16 and S18 have pretty clear delineations, but you'd probably have to include 5 squads in S17.

[Jones]

Thanks for putting these and the rest together