How the Redline Ranking works

Algorithm 1.0

The Redline Ranking orders the squads of competitive REDSEC within each region by their results over the last 365 days. We build it from the placements and kills on this site and nothing else: no votes, no panel, no hand adjustments. It tells you who is playing best right now, and you can check every figure. Each squad's points come with a breakdown that lists every result behind them and what it adds, so you can redo any number here with a calculator.

How it works

A squad's rating averages how well it scored in each game it played. Four factors set how much a game weighs, one adjusts the score for the lobby, and we publish the result as points.

The score: a share of the points available

We score each game under the scoring rule of its stage: placement points plus kill points. Then we divide by the most anyone could have scored in that lobby, a win plus a kill on every player of every other squad. That gives a share between 0 and 1.

Dividing by the maximum lets an Elite Series lobby sit beside an Open Series lobby. The two use different point tables and lobby sizes, so their raw totals can't be compared. The division also explains the small numbers: a lobby win with a normal kill count scores well under half, because no squad takes every kill.

Partner cups predate our per-game record. We hold only their final standings, so each squad gets one result per partner cup, scored as its share of the winner's points.

Tier

Tier describes the event. The Elite Series weighs 1.00, and the weight falls from there; the table under Weights lists every tier. Partner cups and third-party events share one weight, because the circuit matters more than who ran the cup. We list tier 5 events on the site and give them no weight: their fields and scoring vary too much to hold to one standard.

Recency

A game's weight halves every 90 days. A game from 90 days before the snapshot counts half as much as one from that day, and a game from a year ago counts for little. Games older than 365 days drop out.

Roster continuity

Continuity is the share of today's roster that played the game. A game three of today's four players played weighs three quarters of a full one, and a game none of them played weighs nothing. We rank the players behind a squad name, so a squad that replaced most of its lineup starts close to fresh.

Stage

Stage records where the game sat in its event. The weight is 1 + 0.5 × p2, with p running from 0 in the opening stage to 1 in the last. Over a four-stage cup that reads 1.00, 1.06, 1.22, 1.50. The rise comes at the end, once the field has thinned. We keep it small: tier already separates big events from small ones, and inside one cup every game has the same rules against fewer squads.

Field strength

Fourth in a lobby of the region's best beats a win over squads that never place. We credit every share for the lobby it came from:

credited = share1 − 0.25 × f

f compares the mean rating of the other squads in the lobby with the region's mean: 0 for an average lobby, +1 for one twice as strong, −1 for a lobby of zero-rated squads, and it stays inside that range. The exponent keeps the credit on the scale. A share of 0 stays 0 and a share of 1 stays 1, so no lobby turns a wipe into points or lifts a result above a perfect one.

The region's mean counts squads that have played games we hold one by one. A squad we know only from a partner cup has a share of the winner's points, several times larger than the same finish as a share of a lobby, and averaging it in made every lobby in that region look weaker than it was.

The lobby ratings come from the same calculation, so we solve it in passes. Every lobby starts as average. Each pass re-credits every result against the ratings from the pass before, and we stop once no rating moves by more than 1e-9, or after 25 passes. Squads with too few games to rank still count towards the lobbies they played in.

Rating and points

The rating averages the credited shares by weight:

rating = Σ(weight × credited) ÷ Σ weight, where weight = tier × recency × continuity × stage.

Points put the rating on a fixed scale:

points = rating × depth × 3000

A squad that won every game it played and took every kill in every lobby would score 3000. Nobody will. We fix the scale instead of pinning it to the leader, so your points stay put when a squad you never met improves, and a clear leader reads differently from a narrow one.

Regional depth

Regions never play each other, so we rank each region on its own, and a rating shows where a squad sits among the squads around it. Without a correction, the top of a small region would read like the top of a large one. Depth scales a region's points by how much field it has:

depth = √(squads ÷ squads in the deepest region), with a floor of 0.5

A squad counts towards its region's depth once it has 4 counting results in the window. We take the square root because twice the squads doesn't make a region twice as hard to win, and the floor keeps a thin region from reading as worthless. Depth multiplies every squad in a region by the same amount, so the order inside a region stays the same. It matters when we pool regions into one table.

Who gets ranked

A squad needs 10 counting games in the window. A cup runs four games a round, so that means a deep run in at least one cup. Below that we leave a squad unranked instead of ranking it on thin evidence. We will raise the figure as the circuit plays more.

We recompute the rankings once a day and keep every run as a snapshot, tagged with the algorithm version that produced it. Movement arrows compare each squad with the region's previous snapshot.

Weights

Tier	Weight
Tier 1 · Elite	1.00
Tier 2 · Open	0.50
Tier 3 · Partner cup	0.15
Tier 4 · Third-party	0.15
Tier 5 · Unranked	0.00 not counted

Game age	Recency
0 days	1.000
30 days	0.794
90 days	0.500
180 days	0.250
365 days	0.060

Stage of a four-stage cup	Weight
Opening stage	1.00
Stage 2	1.06
Stage 3	1.22
Final	1.50

Share, and the lobby it was taken in	Credited
0.40 in a lobby twice as strong as the region (f +1)	0.50
0.40 in an average lobby (f +0)	0.40
0.40 in a lobby of squads rated zero (f -1)	0.32
1.00 in a lobby of squads rated zero (f -1)	1.00
0.00 in a lobby twice as strong as the region (f +1)	0.00

Squads against the deepest region	Depth
93 of 93	1.000
88 of 93	0.973
46 of 93	0.703
20 of 93	0.500

Region, latest snapshot	Ranked	Depth
AMERICAS 2026-10-05	65	1.000
EMEA 2026-10-05	75	0.962
APAC 2026-10-05	41	0.655

Constant	Value
Window	365 days
Recency half-life	90 days
Games to be ranked	10
Results to count towards depth	4
Stage gain, curve	0.5, 2
Field ease	0.25
Depth floor	0.5
Points scale	3000

Worked example

Team Dragons leads AMERICAS, the region we publish at full depth, as of 2026-10-05, with 47 counting games. We work through the game that adds the most to its points.

In Cash Cup #3 Finals · Lobby 7 · G2 on 2026-10-03 it finished 1 / 24 with 39 kills, scoring 61.0 of 114.0 : a share of 0.535 .

The full breakdown for Team Dragons lists every game. Each ranked squad has one; open it from the squad's points on Teams or from its own page.

Limitations

• No results across regions. Squads play inside their own region, so depth stands in for difficulty. It measures how much field a squad had to beat, which is all the record can tell us. Read the pooled world order as a comparison between regions that never met.
• It knows what we recorded. A game missing from the site is missing from the ranking. Results reach us by hand, sometimes late.
• Kills and placement form one score. The ranking takes each event's scoring rule as it stands, fighting and surviving together.
• Partner cups count once. We hold only their final standings, so a partner cup counts as a single result and can't carry a squad to the minimum on its own.
• A roster with no record reads as continuous. If nobody entered a squad's lineup, we count every game at full continuity instead of dropping it for missing data.
• Early days. The circuit is young and the minimum is low, so a handful of games can move a squad a long way.

FAQ

How can a squad with fewer wins sit above one with more?

Wins are one part of the score. Kills score alongside placement, recent games weigh more than old ones, and a win over a weak lobby earns less credit. The breakdown shows which games made the difference.

My squad didn't play. Why did it move?

Its games aged and weigh less now, or a squad it shared a lobby with played and changed rating, which changes that lobby's field credit. A squad that shared no lobby with anyone who played stays where it was.

Why isn't my squad ranked?

It has fewer than 10 counting games in the last 365 days, its results are all tier 5, or nobody on its current roster played them.

Does signing up for a cup count?

No. Games played count, and a signup says nothing about how a squad plays.

Why does a breakdown sit one point off the published points?

A snapshot stores each rating to three decimal places, and we publish points from that. A breakdown recomputes the rating in full, so the two can round a point apart. A larger gap means someone changed the record after the snapshot, and the breakdown says so.

Can the formula change?

Yes. Any change to a weight gets a new version number. We store the version on every snapshot and leave past snapshots as they were.

Version history

Version	What changed
1.0	First published version. Shares of the available points, weighted by tier, recency, roster continuity and stage. Each result earns credit for the strength of its lobby, solved in passes, and points sit on a fixed scale with a regional depth factor.
wp-1.x	The development line: a week of working out the formula in public. We never published it as a ranking, and its snapshots don't compare with 1.0.