Today we are launching a new stats page on all player profiles that will let you dive into your chosen player's performance with a set of six bespoke data visualizations.
The new page can be found on any player's stats profile right after the overview in a tab called Advanced, which has replaced the Individual tab.

After upgrading our style-based Attributes, we have put the same effort into upgrading our performance-based metrics too. In doing so, we have unashamedly targeted users who always want a bit of extra information.
So without further ado, join us on a journey through the new page.
Advanced Stats Tab summary
Summary: New radar to compare player performance against role averages or any other player
Dueling: A section for eco adjustment, showing a player's performance against their economic expectations and role average
Impact: A section for round swing showing how a player reaches their total value
Round to Round: An interactive density chart showing how many rounds a player has above certain round swing thresholds.
Map to Map: A swarm plot showing all of a player's map ratings inside the sample selected. A box plot showing their floor (C game), median (B game) and ceiling (A game) is on by default
Against the field: A scatterplot showing your selected player against other players in round swing (impact) and eco swing (dueling)
The first thing you will see is a line familiar to all of you from other stats pages. This is where you can select the side and pick your role comparison point. The default will be the player's assigned role; if you're on Danil "donk" Kryshkovets's profile, all statistics you see are compared to the rifler's average. You also have access to the filter bar above, so you can select event type, opponent rank, or a particular group of maps.
If you're looking at a low sample size, a red triangle will appear to warn you that some stats might need more data to be fully reliable.

Summary Radar
Then we see the summary radar chart. This visualization will be familiar to those of you who frequently read EVP articles or our data deep dives, but for those who do not, know that a small circle means a weak performance relative to the player's role, and a larger one means a good one.
The centre of the circle is the 2nd percentile (in an ordered row of 100 players, we take the value of the second worst player), and the outer bound is the 98th (the second best). When a player exceeds those bounds, we display a little triangle at the end of the axis to show they are even better than our chart's usual maximum.
The metrics selected are split into three groups as we go clockwise: Output (Swing Added, eKPR, eADR), Round-to-Round Consistency (Explosive Rounds, Impact Rounds, eKAST) and Differentials (Net Kills, Eco Swing, Round Swing), with Rating 3.0 at the top. These are a collection of our most trusted performance metrics. If you need an explanation, all of them have tooltips.
While some players like Mathieu "ZywOo" Herbaut will fill the entire radar, you can get a sense of how a player gets their numbers when looking at more mortal players.
For example, Drin "makazze" Shaqiri's chart skews towards the right side because he is a high-output player who does not prioritize survival. Yuri "yuurih" Santos, by contrast, skews left because he does not engage in many duels, but does okay when he is part of them.




You can see a carousel of five selected players here, showcasing the triangles for the heights of ZywOo and the lows of Finn "karrigan" Andersen.
You can also compare against anyone you like. On a first click, you are shown the ten players who are most similar stylistically from the doppelganger section on overview as a good starting point. Just remember that if you compare across roles, it's best to switch back to 'compared to: everyone' on the top panel.
Dueling and Eco-Adjustment
The second section, Dueling, is all about pure firepower. Here we are just concerned with whether a player is getting as many kills as their team's investment in them warrants.
Eco-adjustment is the process of how we turn raw kills into something more useful for rating. It is based on win percentages in duels (a duel being any kill or death, excluding team kills), like how an AWP wins 60.7% of direct duels against a rifle on CT side. Eco-adjusted kills are then transformed so that their average is the same as raw kills per round.
But in this tab, we wanted to strip it back before that transformation. By looking at a player's equipment compared to their opponent's, we can calculate their economically expected duel win rate. If you consistently have a better weapon than your opponent, your percentage will exceed 50%, like for AWPers or for star riflers. If you're often on an SMG, or you're looking at a small sample where the player was on eco a lot, it will drop below 50%.
We can then compare this expected duel rate to the actual duel rate (kills divided by total duels entered), the same way we do opening success rate in other stats pages. The difference between these two values is 'Eco Swing', which is highly correlated with eco-adjusted net kills.
Expected duel win rate is our first style stat on the page, so it's the first bar you see that is the same blue as in HLTV Attributes. Every single bar you see on this page is compared to the role average across a similar sample of rounds to what you have selected. In performance stats like eco swing, green is good, orange is average and red is bad.
Below the headline figures, we can show a player's kills, damage and deaths compared to what we would economically expect them to hit. We've also added a tick for the player's role average, so you can compare a player's kills per round against the role average and the economic average in one chart. This is particularly noticeable on an in-game leader.

For example, Benjamin "blameF" Bremer is much more of a star player than usual in-game leaders, and you can see that in the gap between his role average and his economic expectation. But you can also see his prowess in that his actual KPR is way beyond both benchmarks.
Below the dumbbells, we have a few more stylistic or contextual metrics, like a player's average weapon value or how often they get the hero gun on their team. We've also added headshot percentage, since a few people missed it from the overview tab.
Impact and Round Swing
Then we get into the Impact section. This is all about round swing, shown by the big number on arrival.
Next to the overall round swing, we show Added and Deducted. Think of this like seeing Kills and Deaths: The two numbers add up to our overall swing; one is the sum of positive actions like kills, damage, and flash assists, and the other is the sum of negative ones, like deaths.
Round Swing is all about context; a kill matters more depending on whether it's a 5v5 or 4v2, whether the bomb is down, and so on. This can reward different players than eco swing, where here we are looking at the direct impact of a player's duels rather than just whether they won them.

To dig deeper, you can see a stacked bar chart of these two things just below. This shows how much Swing a player receives from each sub-group that goes into it: Deaths, Kills, Damage, Flashes, Traded deaths and team adjustment. When you hover, you can see exact values.
Just below, we have swing stats in certain situations. For example: Clawback Swing, where you can see the round swing of a player as well as their KPR and DPR when their team is at least one player down. There are also boxes for before the bomb was planted, and after it.
Round-to-Round Consistency
Next up, we look at a player's round-to-round consistency. This is important because a player who gets one kill a round five rounds in a row is treated the same as one who scores an ace in one round and nothing in the next four in most metrics we see.
At the top, we have five simple round-to-round metrics like KAST and multi-kill rounds. But below, we have something a bit more exciting.
This is a density plot, showing the distribution of a player's rounds by round swing. On the right, you'll see the percentage of rounds above 0% swing. On the left, the figure shows rounds below 0%. It doesn't add up to 100% because in some rounds a player receives exactly 0% swing.
But it's more than just this. When you hover over the chart, your mouse decides the new zero line. So if you want to see a player's number of rounds above 20%, put your mouse around the 20% mark and the numbers will dynamically change.
This helps show how explosive a player is, as you can see how often a player wins (or loses) a round completely on their own by setting the marker as high or low as you want.
Take a look at Mareks "YEKINDAR" Gaļinskis here, who spikes at -20% (because of frequent impactful deaths), is very low at 0% (very few rounds where he does nothing at all) and is above the rifler average as the chart tails down, because of so many impactful kills.

Map-to-Map Consistency
Moving forward, we look at map-to-map consistency. The five stat bars here relate to how many maps a player has above a certain threshold, and also their rating in wins and losses.
Then there's the swarm plot, coloured by performance. Each dot is a rating on a single map, and when you hover, you can see which game it is from. A click then takes you to that match page.
You'll see three lines by default, relating to a player's C game (their floor, or 25th percentile), B game (their median average, 50th percentile), and A game (their ceiling, or 75th percentile).
If you switch to comparing against the typical role, these values change, so you can see your player's performance relative to the usual in-game leader, AWPer or rifler.
It's a good way to check for consistency. For example, Andrey "tN1R" Tatarinovich can be a match winner for Spirit on his good days but his floor (C game) is only 0.84, and he averages a 0.81 rating in maps his team loses.

Against the Field
Last but not least, we can compare our player against the field more obviously.
This is a scatterplot of our two favoured advanced metrics, round swing and eco swing. Each dot is a different player, and by default you'll only see players with the same role but you can click on the legend to activate other ones.
Along the X-axis, we plot a player's ability to win their duels. Along the Y-axis, it's about their actual impact on the round. Naturally, there's a lot of correlation here, but where a player sits is still insightful.
If a player is in the top left, their kills come at crucial times, but they might not be amazing at winning 50-50 fights. The opposite is true in the bottom right: a good duelist but someone who might make mistakes in high-stakes duels like clutches or opening duels. In the top right, we have the best of the best: the donk and ZywOo corner.

To coincide with the launch of this advanced page, we are also updating what you can see on player stats tables.
Before, you could only see total K-D Differential, K/D ratio and Rating 3.0. Now, we are adding Eco Swing and Round Swing to tables that only include CS2 matches, our two headline metrics from this new page.
On CS2 pages, we are also removing the two KD stats in favour of a better one, namely net kills per 24 rounds. This keeps a simpler metric to judge players by, unadjusted by economy or round context, but in a better way than overall differential (which ignores how many rounds a player played) and K/D ratio (which is more easily inflated by a low death count than differential per round). You can still see all-time kill differential here on the leaderboards page.
We appreciate that this is a lot of data to take on, but take your time, use the tooltips, and before long we're sure that the page will be your portal into judging a player more fairly.
Andrey 'tN1R' TatarinovichPast 3 months
Drin 'makazze' ShaqiriPast 3 months
Mathieu 'ZywOo' HerbautPast 3 months
Spirit
Danil 'donk' Kryshkovets
Andrey 'tN1R' Tatarinovich
Dmitry 'sh1ro' Sokolov
Boris 'magixx' Vorobiev
Myroslav 'zont1x' Plakhotia
Sergey 'hally' Shavaev
Yuri 'yuurih' SantosPast 3 months
Mareks 'YEKINDAR' GaļinskisPast 3 months
Benjamin 'blameF' BremerPast 3 months
Finn 'karrigan' AndersenPast 3 months
Danil 'donk' KryshkovetsPast 3 months

