TSM’s MSI Catastrophe: A Statistical Investigation

Team Solomid’s (TSM) strategy of using Marcus “Dyrus” Hill as a scapegoat backfired during the Mid-Season Invitational (MSI), where opposing teams would gain huge advantages over top lane, despite being even or better in other areas of the map. How did TSM get away with this play-style for so long without being punished? This article is a statistical investigation of Dyrus’s and TSM’s performances over the regular season of Spring 2015, the playoffs of Spring 2015, and MSI.

Individual Play

Here we can see clearly that during the regular season and playoffs, Dyrus’s individual play in terms of his CS mechanics weren’t a hindrance. In wins you could count on him to be at least 5 CS up on his lane opponent, and in losses he was more or less even.

Win Percentage Correlated with Dyrus’s Deaths

Note: If Dyrus was on Sion and had 1 death, it is counted as 0 deaths (because he did a level 1 camp in each of these games)

In the regular season, in games where Dyrus didn’t die within the first 10 minutes of the game, TSM were 11.1% more favored to win the game than overall. However, if Dyrus did die, then TSM were 22.2% less favored to win the game than overall. TSM’s strategy has always been to leave Dyrus top to fend for his own and take advantages elsewhere. If Dyrus was able to hold his own, this strategy would net them huge advantages as the game went on. However, come MSI, teams were able to exploit this hole in TSM’s strategy, and the effect was two-fold.

1. Dyrus was simply not able to handle the pressure in MSI as well as he did during the regular season and gave up gold to the enemy, and;
2. TSM did not gain advantages on the other side of the map as compensation for their allocation of resources.

In the SKT vs TSM game, Santorin attempted two ganks on separate occasions on the SKT duo lane before 10 minutes, the first trading Thresh flash for Urgot flash, and the second gank being completely null. Later, an aggressive invade had Santorin caught out and killed, and combined with Hecarim’s two early deaths to Rek’sai, the score had become 0 to 3.

Win Percentage Correlated with Lane Swaps

Historically, one of TSM’s strengths has been considered to be in their dominant laning. However, in the regular season, TSM was actually 90% likely to win the game if it was a lane swap. If the game was standard lanes (or standard lanes reversed, with top laners in bot lane and bot laners in top lane), TSM was only 62.5% likely to win the game. This is not because TSM is a team that gains advantages for their top laner in these lane swaps, such as teams like CLG and C9 can boast. Dyrus’s regular season CS differentials at 10 minutes is 3.5 CS up for lane swaps and 3.6 CS up for standard lanes; absolutely no difference. Then why do they win more lane swaps?

Again, this goes back to TSM accruing advantages on the other side of the map while leaving Dyrus on his own. When teams send resources at Dyrus, it frees up Bjergsen to play aggressive to gain his CS advantages, and it frees up Santorin to help the duo lane deny the enemy top laner, take the tier 1 tower, and/or take dragon. In TSM’s 8 playoff games, Dyrus played 3 Lulu games, 2 Sion games, and 3 Maokai games. These are champions that require little to no jungle pressure and are incredibly useful as the game goes on. It lets Dyrus thrive in this situation where he received little to no resources and delayed teams exploiting TSM.

During MSI, however, Hecarim was a priority pick for every team, and in order for teams to succeed with carry top laners they need to apply jungle pressure. In the two games where TSM got Hecarim, Dyrus fell flat on his face each time in the face of enemy jungle pressure, contributing hugely to TSM’s losses. In SKT vs TSM, Dyrus’s Hecarim died twice within the first five minutes of the game, and in TSM vs AHQ, Dyrus’s Hecarim died once around the 8 minute mark.

A Strategic Disaster: TSM vs EDG

TSM was hugely off their game, strategy-wise, at MSI. In TSM vs FNC, they did not lane swap the hugely unfavorable Gnar vs Cassiopeia lane, causing Dyrus to get ganked repeatedly and fall behind in CS.

In TSM vs EDG, Dyrus’s Sion died a total of three times within the first 7 minutes, all to dives by Rek’sai, Kalista, and Thresh. In lane swap scenarios, when a duo lane pushes a wave of experience to the tower and the top laner goes to collect it, if there is any danger of a dive, the correct response is to either send the jungler or support to the same lane as the top laner, or your duo lane must dive the enemy top laner himself, or else the top laner simply cannot go collect the experience. Not only did TSM fail to implement any of the previously mentioned countermeasures, they didn’t even send the jungler to at least fast push the duo lane and deny the Mundo. In a lane swap TSM themselves had initiated, they failed even to deny the enemy top laner experience.

Conclusion

In summary, during the regular season TSM was able to gain advantages through mid lane and bot lane while leaving Dyrus out to dry, allowing the team to become hugely successful. During the playoffs, Dyrus was permitted to play Lulu, Maokai, and Sion in every game, allowing him to be hugely effective with minimal support. During MSI, TSM’s weakness had been attacked in multiple direction: Dyrus played champions that were rendered ineffective against enemy jungle pressure; Bjergsen was not able to accrue sufficient advantages against his lane opponent to carry the game (nor was he given the means to in terms of champion picks); TSM’s bot lane were not able to gain advantages despite sufficient jungle pressure. All of these factors above conglomerated together and resulted in a spectacular catastrophe during MSI.