The Deep Dive: Playoffs, Round 2

Originally published at:

How did MAD upset G2?

In a spring split that has been full of unexpected twists and turns and shocking outcomes, the MAD Lions’ upset of G2 Esports in the LEC Spring Quarterfinals may have been the most surprising yet. The young team unseated the World Championship Finalists, knocking them into the lower bracket through a tense, bloody, dramatic five-game affair.

Let’s take a statistical look at how the MAD Lions pulled it off.

Setting the Pace

The biggest factor in MAD’s victory was the way they stepped up to play at G2’s pace, leading to a drastic increase in their combined kills per minute (CKPM) relative to the regular season. The series saw an average of 1.13 champion kills per minute between the two teams, a small increase for G2 but a jump of 55% for MAD.

Regular Season 0.73 1.10 3.34
Quarterfinals 1.13 1.22 3.72
G2 Esports CKPM K:D WPM
Regular Season 1.01 1.61 3.40
Quarterfinals 1.13 0.81 3.66

2020 Spring Split

One explanation for this shift in play style may be that MAD were playing from the comfort of their own homes, rather than sitting on a stage in front of fans where they might feel more nervous and play more conservatively. Another reason may be that the coaches did a good job of preparing the players and intentionally increasing their comfort level. Or maybe MAD felt like they had nothing to lose, since they were such heavy underdogs, so they had no pressure and were able to throw caution to the wind.

From a strategic perspective, the MAD Lions’ willingness to embrace G2’s aggressive “skillcheck” style was very intelligent, and if the coaches encouraged the team to play this way for strategic reasons then I applaud them for it. This young roster was put together for its raw skill and potential, not for its experience and measured decision-making. If the MAD Lions attempted to slow the game down, G2’s greater cumulative game knowledge could have easily overwhelmed them. To put it another way, G2 were probably 75% favourites in a macro matchup, but MAD have a very skilled roster with rough edges, so in a brute force skill matchup, G2 may only have been 60% favourites.

It’s a tactic as old as time: if your opponents are the “better” team, increase your odds of an upset by introducing more variance into the game. Typically the best ways to add variance are a) playing off-meta picks and b) playing aggressively and taking more “50/50” plays. MAD took option B and rode it to victory, while G2 were actually the ones playing further off the meta with their Kalista top and Ziggs/Bard bot lane and some other “fancy” picks.

Another noteworthy change MAD made, statistically speaking, was upping their vision coverage, increasing from 3.34 wards per minute (WPM) in the regular season to 3.72 in the Quarterfinals. G2 also increased their vision coverage from 3.40 WPM to 3.66, but MAD’s increase was more sizeable and made the bigger difference. The greater vision flow—which was mostly driven by Kaiser’s boost from 1.54 WPM in regular season to 1.96 in playoffs—gave them better information to find the skirmishes and team fights they wanted.

Objective Control

The series wasn’t all about combat. There were also some interesting trends in neutral objective control.

MAD Lions Neutral Objective Control

Objective Regular Season Playoffs
Rift Herald 66% 70%
Dragons 60% 67%
Barons 71% 63%

2020 Spring Split

MAD Lions took 7 out of 10 Rift Heralds (70%), 16 out of 24 dragons (67%), and 5 of 8 Barons (63%) in their match against G2 Esports. That was a great continuation from the regular season, and better than might be expected against a team as good as G2.

Then again, G2’s neutral control wasn’t as impressive in the regular season as you might expect.

G2 Esports Neutral Objective Control

Objective Regular Season Playoffs
Rift Herald 55% 30%
Dragons 69% 33%
Barons 53% 37%

2020 Spring Split

Giving up Rift Heralds and dragons certainly cost G2, but they barely took more than half of the Rift Heralds and Barons in the regular season. That suggests that G2 has either undervalued Rift Herald in general this split, or done a poor job of executing on control of that objective. Given G2’s league-leading +1,401 gold lead at 15 minutes in the regular season, it’s unlikely that G2 didn’t have opportunities to take more Heralds. G2 may need to re-evaluate the way they play to the top half of the map.

MAD Lions Jungle Carry

Shad0w deserves a special call-out for his performance in the Quarterfinals. He increased his KDA, farming, and damage output, taking on a more front-and-center role.

Regular Season 3.6 5.0 291
Quarterfinals 6.3 5.6 430*

2020 Spring Split
* Games 2-5; damage data unavailable for game 1

Shad0w’s game on Kindred probably helped increase his CSPM and DPM, but his champion pool overall was fairly standard. Along with the Kindred, he played Lee Sin twice and Gragas and Elise once each.

The real change was that Shad0w upped his aggression and selfishness and made good use of both of those “resources”, delivering great performances with a high proportion of successful plays and a huge increase in damage output.


The road from here doesn’t get any less challenging for Shad0w and his teammates, with Fnatic on Saturday and potentially a runback against G2 afterwards, even if they win and make it to the Finals. MAD narrowly won this series, and it was a huge accomplishment, but G2 will probably come back stronger.

MAD now need to sustain the good work they did last weekend, and further refine it. They’ve proven their ceiling, and if they can maintain consistency then nothing will be out of reach.