Why you shouldn't judge League of Legends pros based only on their KDA

KDA ratio is the best-known and most widely-used statistic in League of Legends. But while it shows us how many fights a player wins, it isn’t a good enough stat to stand alone. That’s because KDA is heavily influenced by deaths, but dying isn’t always bad.

To accurately gauge player performance, we need to use stats properly, and that means making sure we’re measuring the right things, with the right context. Let’s dig in with an example: here are the KDA ratios of three NA LCS mid laners, following Week 8 of the 2017 Spring Split. Based solely on these numbers, Bjergsen has been the strongest performer by far, while Froggen has been a little above average and Hai is way out of the running. But this isn’t the whole story, and we’ll show you why. It’s time to expand our horizons with some core stats that every aspiring League analyst should have in their arsenal.

First up, here are some more nuanced measurements around kills and deaths. Kill participation, or KP, is calculated as a player’s kills plus assists, divided by the team’s total kills. Death share works the same way, but with… deaths. KP tells us how much a player contributes to their team’s fights, while death share tells us how well the player balances their involvement with survivability. Quick note: when comparing between players, remember that a higher KP isn’t always better, and a higher death share isn’t always worse. For example, some teams play a more spread-out style that leads to lower KPs across the board, and some deaths are more worthwhile than others.

KP and death share are useful stats for all roles, but KP is especially meaningful for measuring the impact of mobile playmaking roles like jungle and support, while death share is most telling for mid and AD carry positioning, as they’re often priority targets in fights. Here are the kill participations and death shares for our sample mid laners. Although Bjergsen has the highest KDA in this group, Froggen actually leads in KP and death share, showing just how crucial he has been to Echo Fox’s success. Hai has the highest KP on FlyQuest, but his death share is ugly.

This tells us two things: first, he takes too many risks trying to create plays for his team, and second, he’s often targeted by his team’s opponents. It would be great if Hai could tighten up his play and keep his death share under twenty percent while still leading his team in kill participation. Next up are two common resource metrics. Creep score per minute is a player’s last hits on minions or neutral monsters, divided by game length.

Gold share is a player’s earned gold divided by the entire team’s earned gold, after removing starting gold and inherent gold generation. These stats help us understand a player’s role on their team: higher resource numbers should raise our expectations for fight involvement and damage output, while lower resource numbers suggest a more supportive role. Bear in mind that CS per minute and gold share are affected by both a team’s willingness to funnel resources into a player and by the player’s ability to generate resources for themselves. Resources are both distributed and contributed, so context is key. Here we again see Hai trailing the others, picking up far less farm than Froggen or Bjergsen.

Froggen has a larger gold share than Bjergsen, and Hai has the same gold share despite a much lower CSPM. That suggests either that they’re both out-earning their teammates in kill gold, or that their teammates are worse than Bjergsen’s at farming. Hai leads FlyQuest in both kills and farm, so while we can’t necessarily praise him because his numbers are still lower than the average, we can say that his teammates are letting him down in the side lanes. As for Froggen, these numbers confirm just how heavy a load he’s carrying: his ADC has the lowest CS per minute in the league at his position, and his top laner is second-to-last. Compared to Bjergsen, both Hai and Froggen play more central roles for their team, so we should evaluate them with that in mind. Now we hit the really fun stuff: damage to champions.

Everyone wants to know how hard their favorite player is carrying, but be extra careful: damage stats are the most context-heavy numbers in League of Legends. Damage is very heavily influenced by champion choice and game length: poke champions like Corki, Jayce, and Varus will always produce high damage totals, and the longer a game goes, the more the per-minute numbers will get inflated as they level up and buy bigger items. When possible, it’s best to look at damage per minute, or DPM, and damage share together. For example, a high DPM with low damage share means the whole team’s DPMs are probably inflated by game length and frequent fighting.

Damage stats are most useful in determining the impact of carry roles like Mid and ADC, who are often the main damage dealers. Before we read into the numbers for our mid laners, let’s identify potential champion biases: Hai has played ten games on poke champions, while Froggen and Bjergsen have each put in six. We should expect to see a little bit of inflation on Hai’s damage output. Sure enough, look at Hai’s numbers: they’re way up there! In fact, Hai and Froggen have the two highest DPM marks in the entire NA LCS. Hai is definitely carrying his team, and he’s doing it with less gold in his pockets than Froggen, even if his champion pool is boosting his numbers a little.

You may be surprised to see how much lower Bjergsen’s damage numbers are, even though he’s played just as many poke champions as Froggen and gets more farm. Bjergsen gets a little leeway because his team fights less often than Froggen’s, but it’s fair to say that Froggen is generating more value from his gold than Bjergsen, as far as damage goes. Time to put it all together: we started with a clear tier list based on KDA ratio, with Bjergsen leading and Hai way behind. Kill participation and death share brought Froggen closer to Bjergsen’s level.

Some resource numbers helped us understand how central Hai and Froggen are to their teams’ success. Finally, we looked at damage output, where Froggen and Hai both came out looking pretty good. Based on all of these stats, Froggen actually looks like one of the best performing mid laners in North America, while Hai is providing a lot more value than his basement-tier KDA might suggest. The journey doesn’t end here. There are more stats to consider, from laning to warding.

And of course, we need to watch the games closely to understand where the numbers are coming from. But armed with this information, we can have a better idea of what to look for: Why does Hai die so often? Why does Froggen get so much of his team’s gold?

Why are Bjergsen’s damage numbers so low? Better focus makes for better analysis.