Introducing Rating 3.0

Economy adjustment and Round Swing, a metric that measures the direct impact of each kill, are live today on statistics and match pages.

Rating 3.0 has been released across all Counter-Strike 2 matches in rating's biggest overhaul since 2017's 2.0.

The rating is made up of two parts, a version of rating 2.1 that is adjusted on economic factors supplemented by a brand new Round Swing metric.

We now have six crucial sub-ratings, rather than 2.1's five: Kills, Damage, Survival, KAST, Multi-Kills, and Round Swing.

Like previous ratings, the average is 1.00 over a CS2 event, and we'll use this article to explain how it works, why we made the decisions we did, and at the end show you who is most affected by the swap.

Changelog

• Round Swing added, a metric that looks at a team's win probability before and after each kill with knowledge of the map, side, and economy of each round.

• Credit for the swing of each kill is divided based on the final point of damage, the damage share, flash assists, and whether it was a trade kill.

• Impact replaced by a combination of Round Swing and Multi-Kill rating

• Kill, Damage, Survival, ratings are all eco-adjusted based on the win-rate of a duel based on the map, side, and equipment of the individuals involved.

• KAST and Multi-Kill ratings are eco-adjusted based on the probability of a given action taking place given the map, side, equipment of the individual, and average equipment value of the opposing team.

• Assists reset to CS:GO value of 40 damage for all purposes.

• We have added the concept of trade denials (two kills within 5 seconds), and failed trades (punishing the second player, and rewarding the first, if both die within 5 seconds) within the sub-ratings.

Where you'll see Rating 3.0

The most obvious place to start is on players' stats page. Here is an example of Mathieu "⁠ZywOo⁠" Herbaut in 2025.

The design has been refreshed for 3.0, with T and CT rating visible alongside rating at the top. There is also a splash of team color behind a player's bodyshot on the left.

Multi-Kill rating has gone where Impact was, while a new space has been added for Round Swing. We also show you the maps in the sample in a more obvious way than lower down in the page or sidebar.

The red-yellow-green bars underneath each metric relate to a player's sub-rating in that category, not the metric you can see. This was true in 2.0 and 2.1, but with eco-adjustment it may happen more often that a player has a higher raw figure, say 0.70 KPR but a yellow bar, versus a player with a green bar and 0.68 KPR.

A toggle has been added to show eco-adjusted figures that will align more closely with the color of the bars, but remember that far more goes into these sub-ratings than just the main figure you see.

This graphic shows just how many metrics are put into each sub-rating before a number (and the color of the bar) is spat out on the other end.

The other main place you will see rating 3.0 is match and result pages. When a match ends, you will be able to see swing and 3.0 on the scoreboard. We have also added openers, multis, and clutches to match result box scores.

Everywhere else on the site you expect to see the constituent parts of rating, like the 'performance' tab on result stats pages, player of the match or MVP boxes, and highlighted stats on match pages, will show you both multi-kill rating and Round Swing.

Later on, we will integrate eco-adjustment and round swing into the attributes section to refine the stats page further.

Economy adjustment

This is the big feature you have all been asking for. A key flaw of previous ratings was that each kill was missing the context of if it was performed against an armorless, Glock-wielding opponent or a kitted-out AWPer.

The calculation is based on the price of a player's armor plus their most expensive weapon, leaving groups of what we can roughly call: Sniper, tier-one rifles, tier-two rifles, SMGs and shotguns, upgraded pistols, and starter pistols.

Only 44.7% of duels in Counter-Strike take place between players in the same economy category. For more than half of the kills that take place in a game, therefore, there is an economy imbalance.

Using those categories, we can calculate the win percentage for each economy group against one another. Like so:

Intuitively, the more expensive your weaponry (and this just counts armor and your weapon) the better chance you have of winning a duel.

In Rating 3.0, these figures are crunched to leave you with an eco-adjusted version of a kill that is directly based on statistical fact. If a duel is harder to win, you get more kill points. If it is easier, you get fewer.

The primary goal of this process is to reduce the impact of eco frags on rating. Killing armorless players, both here and in Round Swing, gives you far less rating than kills in gun rounds where a duel is more 50-50. For example, on T side, a rifle vs rifle kill (48% win rate) counts as around 1.10 'kill points.' Killing a starter pistol (75% win rate) drops that figure to 0.54.

Since pistols are powerful in the game, and duel-win rates are pretty high even in the last category, do not expect eco frags to have no impact at all. But they have been mightily reduced. The vice versa is true too, where kills with low-value equipment on full-buying players gives you more points.

The other consequence of eco-adjustment is a nerf to AWPers, when they wield the Big Green. The one-shot-kill weapon is a staple of every team for a reason, and the reason is visible in the table: AWPers win 56% of their duels against riflers on T side and 60% on CT side.

AWPers were not overrated, per se, in rating 2.0 and 2.1 because of 25 damage assists and their role causing lower Damage and Impact rating. When 2.1 heavily nerfed surviving in lost rounds, passive AWPing was hit again. We will continue to monitor the impact of eco-adjustment but snipers have plenty of space to score highly, by using their gun well in favorable duels and via the one metric we have not got to yet.

When we combine our eco-adjusted metrics, we essentially have an eco-adjusted and improved version of rating 2.1, rating 2.5, you could say. But to make it 3.0, we also have:

Round Swing

Anyone who is familiar with our EVP articles will already be familiar with Round Swing, but here's an explainer for those who are not.

Round Swing looks at each kill and sees how much it changes a team's chance of winning the round. That includes each team's economic situation, whether the bomb is planted, how many players are alive on each team, and which map the player is on for targeted CT-T percentages.

It also looks at more than just which player got the kill. Share of credit is divided based on yes, who got the final point of damage, but also the damage share, flash assists, and if the kill was a trade.

Divided per round, we are given percentages for each player, with the best of the best sitting around +4.0% (Danil "⁠donk⁠" Kryshkovets +3.79%, ZywOo +3.69% in 2025 thus far) for a season and mortals between -1.5% and +1.5% over that large of a sample size. Figures, of course, have greater magnitude over an event or a map.

It's a complicated metric, but it is worthwhile because it knows of the context of each kill. A kill in a 3v3, or 2v2, is more important than in a 5v2 or 4v1. If the opponent is on a full eco, and your win probability is 96%, your team can only gain +4% round swing in sum rather than the +50% of a regular 50-50 round.

Knowing the players alive also means that players are rewarded for saving their team in difficult situations: If the win probability drops due to your teammates dying against an eco, you receive credit for bailing them out. So, rating can still be gained with favorable machinery if the round becomes perilous.

Opening kills are another example of where round swing is more fair. In 2.1, players received more credit for opening kills than punishment for opening deaths, in an attempt to reward aggressive players. In 3.0, round swing can correctly see if a player's opening kills are worthwhile — or if they immediately die and it's a 4v4 (a good trade on T side, but not on CT).

There are some drawbacks to this metric, which is why it is not 100% of rating 3.0. It is a feature, not a bug, that multi-kills have diminishing returns: Turning a 5v5 into a 5v3 is immensely valuable, but turning a 4v2 into a 4v0 gives you much less Round Swing. Statistically this is a valid conclusion but rating is also about crediting players with the ability for those explosive moments, which is why Impact rating has been replaced by both Round Swing and Multi-Kill rating.

Early-round moments, because there is so much to play for still, also tend to lead to less of a reward than late-round ones. An opening kill is around +20%, but winning a one-on-one clutch might be +50%. Winning even harder clutches will be even more, something worth considering over smaller samples of round swing like over a single map.

Statistically again, this makes perfect sense, but for our purposes (measuring the best CS players, which isn't 100% correlated with direct statistical impact on a round), those early rounds need more of a say.

One measure we took was to extend the value of each round if losing players survive based on players' contributions. So, if you secure a double entry as a Terrorist, the round win probability jumps from 48% to 89%. If the three CTs then save their weapons, you are credited for the final 11% too.

In the real world, saves usually happen after multiple Ts have contributed, and in that situation it is divided based on your share of the contribution up to that point. Because Round Swing is zero-sum, the saving players are punished the same as if they had lost that clutch. Your reward for saving is your weapon in the next round, not some extra rating points.

And, as we have said, eco-adjustment is kinder to riflers than previous ratings. Therefore, it is not the end of the world if Round Swing slightly favors more passive players (who tend to be snipers), given the rest of rating is kinder to more aggressive ones.

Where has Impact gone?

A key part of Rating 2.0 and 2.1, Impact is no longer a part of 3.0. Round Swing is its main replacement, in terms of measuring the impact of a player's output, but its spiritual successor is Multi-Kill rating.

Impact was based on opening kills, clutches, and multi-kills, so Round Swing covers most of its function in a more precise fashion. Multi-Kill rating then rewards explosive moments with knowledge of the economy to ensure that element of play is still reflected on the scoreboard.

Who is most affected?

When this is all put together we can build a picture of which stocks are rising and falling. Aggressive riflers still have the highest average rating, but select examples like Dorian "⁠xertioN⁠" Berman, Mario "⁠malbsMd⁠" Samayoa, and Mareks "⁠YEKINDAR⁠" Gaļinskis will drop a good deal after this update.

These are players who are high-activity both in gun rounds but also in anti-eco and anti-force situations, where kills count for less and deaths count for more. Giving away man advantage situations is also punished heavily by round swing.

Add in assists being 40 damage instead of 25, and opening deaths being punished as much as opening kills are rewarded, and aggressive riflers are now on a more even playing field.

Their inverse, the supportive, eco-friendly, rifler is on the up shown by players like clutch king Jimi "⁠Jimpphat⁠" Salo and the selfless Sodbayar "⁠Techno⁠" Munkhbold. These are players who rarely give up their life for free, often have weaker weaponry, and can come up clutch.

AWPers can also go up and down, with Lucas "⁠nqz⁠" Soares rising and Nicolai "⁠device⁠" Reedtz and Usukhbayar "⁠910⁠" Banzragch going the other way. Like with openers, we can see that the average rating stays roughly the same but it is different players that are rewarded.

Mid-round experts like Russel "⁠Twistzz⁠" Van Dulken, who specialize in the 3v3s that decide matches, can expect a boon while high round win percentage players like Dan "⁠apEX⁠" Madesclaire will need to do more to get the same rating as before: It is harder to coast to high ratings off your team in 3.0.

An eco-fragging example

Let's also look at an example scoreboard, with FaZe and Jonathan "⁠EliGE⁠" Jablonowski starring in this performance vs BetBoom.

In rating 2.1, EliGE's 133 ADR and 23 frags pushed him to a 2.06 rating. But, nine of his 23 frags were against full ecos, and an additional three were against half buys. When you eco-adjust that, a good chunk of his rating disappears.

Some of that eco-fragging was important, and he still boasted a good record in equal situations (4-1 K-D in pistols, 84 ADR in gun vs gun), so his rating stays at a respectable 1.40. But it goes a long way down from the 2.06 he had originally, allowing Helvijs "⁠broky⁠" Saukants, who had a solid game and a 1v2 clutch, to squeeze ahead.

Over larger samples, eco fragging evens out for the most part. But in small single-map or series situations, we will be able to isolate a player's true impact more quickly than before, even if it will take a while to get used to a player with 133 ADR not always having the most impactful performance.

History of rating

Rating 1.0 was introduced to the site in 2010 by Petar "Tgwri1s" Milovanovic as an improvement to K/D Ratio in evaluating player performance in professional Counter-Strike with multi-kills as a third factor in addition to kills and deaths.

Read more

What is that rating thing in stats?

That core system remained in place for 2017's rating. 2.0 update, which introduced a more complicated formula. New metrics like KAST and Damage, and a reformed Impact rating added opening kills and clutches to the multi-kills that drove the majority of the sub-rating.

Read more

Introducing Rating 2.0

This meant supportive players who might contribute in terms of unfinished kills, but dealt lots of damage, assists, or had their deaths traded, were rewarded. Impact's new formula also gave a hand to aggressive players, for their openers and multi-kill potential.

Rating 2.1 was then released last summer to adjust the formula's averages for Counter-Strike 2, and to punish saving.

Rating has and will continue to guide all statistical decisions at HLTV since its introduction, and is the backbone behind Most Valuable Player (MVP) awards and the prestigious Top 20 players of the year list.