This week we continue our journey down the rabbit hole of teamfights in Overwatch. Previously, I provided the definition of a teamfight and explored some factors that determine teamfight winners. With that context, it’s time to use teamfights to identify “clutch” players in Overwatch. 

To determine who died first in teamfights, I had to create certain filters to bring these eliminations to the forefront. These same principles can be applied to introduce more specificity to eliminations in teamfights and hone in on the final blows that make or break matches. It’s these clutch final blows that we’re going to explore today. (Note: for most of this article I will be using 2018 data, as there is a larger sample size to draw conclusions from.)

What’s the first play that comes to mind when you think about clutch moments in the Overwatch League? For many, it would be something like Jae-Hyeok “Carpe” Lee’s winnable moment. In Overwatch, it’s difficult for individual skill to make a game-winning impact, but we can all agree that Carpe’s several eliminations in a row—all while his team was down players—definitely made a difference. With this in mind, I dove into teamfight data and pulled out only final blows that occurred while a player’s team was down at least one person. These will be referred to as “shorthanded” final blows.

I then converted these shorthanded final blows into a rate over 10 minutes, using the duration of the teamfights they occurred in. The resulting rate is plotted against teamfight win rate* below:

1.png
Shorthanded final blows/10 mins in teamfights plotted against teamfight win rate*. (Min. 150 shorthanded final blows in teamfights in 2018)

*Draws are included and treated as half of a win, half of a loss

A lot of these results match our memories of the inaugural season. Nam-Ju “Striker” Gwon’s ability to pull Boston out of seemingly lost situations was a big factor to their historic 10-0 Stage 3 run. Georgii “Shadowburn” Gushcha—for all the questions that surround him—was always known to be clutch, especially on his signature Genji. Hyeon “Effect” Hwan and Weida “Diya” Lu would often drag their teams kicking and screaming to map wins, their teams having given them plenty of shorthanded opportunities. On the opposite side of the team strength spectrum, fans of the Overwatch League could always agree that Jun-Young “Profit” Park, Jong-Ryeol “Saebyeolbe” Park, and Byung-Sun “Fleta” Kim were clutch, but now they have the numbers as proof.

However, this is only one way of looking at clutch players, and an incomplete one at that. Recall that landing the first elimination in a teamfight tends to lead to a teamfight win over 75% of the time. By definition, these initial eliminations cannot be shorthanded, but surely the players who land these initial final blows are also clutch! It also follows that any final blow that occurs while both teams have equivalent numbers of players would confer a similar advantage to winning the teamfight. When we add these “even-strength” final blows in teamfights to the shorthanded cohort, some new faces begin to appear:

2.png
Shorthanded and even-strength final blows/10 mins in teamfights plotted against teamfight win rate, with draws included. (Min. 300 shorthanded and even-strength final blows in teamfights in 2018)

Suddenly we have Widowmaker and Junkrat players. Many teamfights in 2018 were decided by a clutch initial pick by a flanking Widowmaker or a big Junkrat RIP-Tire on a support player. Players like Do-Hyeon “Pine” Kim, Seok-Woo “Wekeed” Choi, Jeong-Woo “Sayaplayer” Ha, Min-Ho “Architect” Park, and even Jake “Jake” Lyon stand out when including even-strength final blows alongside shorthanded final blows due to their hero pools in 2018.

This is all well and good, but we can go deeper—much deeper.

Let’s revisit Carpe’s winnable moment. What else made it so special? As the clip begins, the game informs us that we’re entering overtime. Overtime is special in Overwatch because it can occur multiple times in a match, and each overtime can result in the end of a round, making it an inherently high-pressure, high-stakes situation for both the attacking and defending team.

What if we only looked at teamfights that either started in, or contained the start of, an overtime? What if we added this context to our population of shorthanded and even-strength final blows? What if we plotted that against a player’s teamfight win rate in overtime? Check it out:

3.png
Shorthanded and even-strength final blows/10 mins in overtime teamfights plotted against teamfight win rate in overtime, with draws included. (Min. 50 shorthanded and even-strength final blows in overtime teamfights in 2018, Stage 3 and later)

Unfortunately, the ability to statistically identify overtime in Overwatch League matches was not implemented until Stage 3, so players like Shadowburn will not show up in this dataset. Sorry! However, this additional context does start to cast light on the most clutch players of 2018. Simon “Snillo” Ekström—during his brief appearances for Philadelphia—always had insane final blow/death ratios on his signature Tracer. Lane “Surefour” Roberts, who greatly improved in the latter half of the season on Widowmaker, stands out with the second-highest rate of these clutch eliminations.

Still, we can go even deeper. Up until now, we’ve only analyzed the rate of final blows in teamfights under tighter and tighter constraints, but we have yet to consider the outcomes of these teamfights. What if we only looked at the rate of clutch final blows in teamfights that these players ended up winning? Now we’re talking. See below:

4.png
Shorthanded and even-strength final blows/10 mins in overtime teamfight wins plotted against teamfight win rate in overtime, with draws included. (Min. 35 shorthanded and even-strength final blows in overtime teamfight wins in 2018, Stage 3 and later)

Because the x-axis has been filtered for teamfight wins only, the y-axis is purely meant as a reference—these players’ teamfight win rates are all 100% under these conditions. Short of creating some sort of rating that assigns extra value to shorthanded kills based on the degree of shorthandedness, this is as close as I can get to identifying the most clutch players in the 2018 season. These players landed final blows that were:

  • While their team was down a player or at the same strength as the opposing team.
  • In teamfights during overtime or containing the beginning of overtime.
  • In teamfights that their team ended up winning.

It doesn’t get much more clutch than that. With that in mind, I hereby crown—drumroll—Architect as the most clutch player in the latter half of the 2018 season.

Looking at this graph, the most interesting thing to me was the separation—or lack thereof—between DPS players on the same team. Architect may have led the league in 2018 in clutchness, but he also led his team by a large margin; Jay “Sinatraa” Won did not even fulfill the minimum of 35 final blows required under this filter. A similarly visible gap appears for Surefour and Joao Pedro “Hydration” Goes Telles—perhaps due to the resources funneled to the Widowmaker player as a force multiplier to his own skill. Also note the gap between Los Angeles Valiant DPS duo, Brady “Agilities” Girardi and Terence “Soon” Tarlier (who was the Widowmaker specialist), even though the latter held the reputation as the clutch player on the team.

The difference in “clutchness” does not bear out for every team, however. Boston’s 2018 DPS duo of Striker and Stanislav “Mistakes” Danilov held nearly equivalent rates of clutch final blows, as did Jake and Jiri “Linkzr” Masalin. Ever frightening, New York’s trio of Saebyeolbe, Pine, and Hae-Seong “Libero” Kim also showed only slight gaps in their rates of clutch final blows.

Whether it’s one clutch player, two, or three on a team, there does not seem to be a definitive way to build a roster around this quality. After all, clutching is not strategic—it’s what we call something that isn't supposed to happen. Having a player that has the ability to clutch when their back is against the wall is valuable, but there’s something to be said for winning games without needing players to unleash their inner Carpe in the first place—although, as London and Philadelphia proved in the end, it’s pretty awesome to have Carpe and Profit anyway.

As a bonus feature, let’s look at the current season’s clutch players, albeit with a much smaller sample size:

5.png
Shorthanded and even-strength final blows/10 mins in overtime teamfight wins plotted against teamfight win rate in overtime, with draws included. (Min. 10 shorthanded and even-strength final blows in overtime teamfight wins in 2019)

Once again, note that the y-axis of overtime teamfight win rate is purely given as a reference. The cross-section of teamfights, overtime, shorthanded or even-strength, and teamfight wins is currently quite low for most players, so we have to take this data with a massive grain of salt. For example, teams like the Uprising and Gladiators have changed their lineups during Stage 1, resulting in some players like Gui-Un “Decay” Jang or Cameron “Fusions” Bosworth not hitting that 10 final blow cutoff.

Still, we can see that clutch final blows tend to skew towards D.Va and Zarya players in the current meta, with some Reinhardt standouts in Young-Jin “Gamsu” Noh and Matthew “Super” DeLisi. Lucas “Note” Meissner’s Self-Destructs, as well as his ability to hunt down supports, have been something we’ve been highlighting on the broadcast, and it’s cool to see that his stats back up the eye test on the screen.

As we look to the Stage 1 Playoffs, which player do you think will end up with the most clutch plays? Will it be Hyun-Woo “Jjanu” Choi, whose Jja-nukes have shocked the world? Will Note continue to lead the Uprising and the league, carrying them to a deep playoff run? Stay tuned...

Ben "CaptainPlanet" Trautman is the statistics producer for the Overwatch League global broadcast. Follow him on Twitter!