Nkashyap Looks At: Na’vi’s ESL Pro League Failures
This is the start of a new series of writing pieces that I hope to put out that will combine my personal experience as a counter strike player and analytical viewer with my life experience as a software engineer in order to drill down with an alternative perspective into the mechanics of professional counter strike.
Just recently, on April 11th 2021 to be specific, ESL’s latest iteration of its counter strike tournament, ESL Pro League Season 13, came to a conclusion when Cadian’s Danish Heroic side beat one of CIS’s best contenders. The surprising fact however was that the CIS contender that finished top two was not actually the CIS team that found itself at the top of HLTV’s rankings at the time of the tournament, no that team in fact failed to get out of the group stages. This result was only the follow up to IEM Katowice, where Na’vi also found themselves outplaced by other competitors from the CIS region. At least then proponents of Natus Vincere could point to the fact that they placed on equivalent footing as Astralis, and subsequently argue that they should not be castigated for shaky results in this turbulent online era. However after I parsed the demos of Na’Vi at ESL Pro League, and began to dive deeper into the actual data behind it I personally slowly began to align with the view that Na’vis failings at ESL Pro League were unacceptable and explicable of deeper seated issues within the roster construction of Na’vi.
S1mple does it all
The first thing that immediately stood out to me upon parsing Na’vi’s matches at ESL Pro league, which consisted of six best of threes versus teams like Mibr, Cloud 9, Complexity and more, was that s1mple stood out as statistically dominant in such a way that lines up with what one witnesses when they spectate s1mple on the server. The first fact that was extremely indicative of the dominance that s1mple exudes every time he steps on the server was the fact that he was providing fragging output for Na’vi throughout almost every location on the map. I parsed all his kills and plotted them giving me the following visualizations.
S1mple kills on Mirage
(Clusters are indicative of a greater volume of kills coming from certain spots)
Perfecto kills on Mirage
Electronic kills on Mirage
Perhaps the cursory glance you took at the visualizations of S1mple’s kills, and also his teammates, did not allow you to glean the same perspective that I did. Let me enlighten you on as to why these three images, from my perspective, illustrates how s1mple is the backbone of Na’vi that causes the engine to go while his teammates can be more viewed as ancillary pieces that fill a variety of roles from supporting stars to even role players.
The reason is that s1mple is impacting the game on a high level all across the map whereas electronic and perfecto are clearly filling roles on the map, albeit at a high level. Look at perfecto’s heatmap, you can clearly see clusters around bench, and van on B site, clearly showing that he’s investing his time on the B site functioning as a B anchor. On the contrary on electronic’s heatmap you can clearly see that a large portion of his impact comes from the connector area, a role that he fills at a high level on CT sides. Now you may say that perhaps s1mple is just a more well rounded player, but that at those specific locations his teammates are performing just as well or if not even better than him. However the fact of the matter is that you would be wrong to say so as when you drill even deeper into the data it becomes even more clearly how statistically dominant and important s1mple is. I also parsed the data and plotted K/D Differentials for na’vi across locations on maps. Nothing complex, simply if you got a kill at a location that is +1 and if you died -1. When looking at locations such as B site on mirage, and even connector on mirage, s1mple came out looking prim and pristine even when compared to his teammates who specialize in those areas.
As you can see S1mple is dominant even in the locations that electronic and Perfecto specialize in.
With all that said however I would take this with a grain of salt as at the end of the day it is just a number stripped of a lot of context of how the games proceed. Perfecto after all is a B site anchor and in those situations it can be hard to maintain a positive kill death differential, and some would argue that it not even be a requisite for good play as seconds bought thorough good utility usage could be just as useful.
Regardless the more you drill into the data the more points that pop up to drive the inquisitive mind towards one inevitable conclusion. That s1mple is a monster who essentially functions as the fragging backbone that Na’vi uses to keep its place at the top end of the scene.
Time for New Blood?
Another characteristic that became more apparent the more I drilled into the data was that flamie was a player who was filling a role, but not necessarily at an adequate level for tier one play. Take for example frags a simple but essential prism to look through the game at, due to the fact that cs go at the end of the day will revolve around clicking heads, or in my case a prism that I scaled to frags per 30 rounds so that we have a standard way of measuring across players.
At initial glance flamie does not seem to have a very problematic frags per 30, however one can account for Perfecto’s and Boombl4’s lower level of fragging output as they fill some of the more utility aspects of the team. Take perfecto on mirage for example, he takes the less sought out role of b anchor, a position that is frequently targetted and punished by elite teams with well thought out utility usage and multi pronged attacks. Similarly Boombl4’s lower level can be chalked up to his simultaneous presence as the team’s in game leader, however there exist less mechanisms to explain away flamie’s lack of presence as a tertiary fragger. It is not like he offers playmaking or any other ancillary benefits either, when taking a look at his kill locations on ct side it becomes quickly clear that he either functions as a site anchor or a retaker if he is alive for that point, he doesn’t really go for flanks or creatively take map control either.
(Flamie on Dust 2 CT)
Furthermore it was not like his utility usage was exemplary as a role player either as can be seen below in the he damage chart.
Two things immediately jump out, one the fact that s1mple manages to create a kill for himself through he usage almost every 30 rounds, and two the fact that flamie has the worst he damage statistics, even when compared directly to the person replacing him B1t.
While statistics can often be used to mislead when stripped of context, when all the numbers point to a key component in the machine underperforming at some point you will be forced into action to remedy the situation. That is exactly why I’m extremely excited for the future of Na’vi, because I believe that the b1t replacement can function as a point of arbitrage that Na’vi can use to reclaim fragging output that was left on the table by Flamie. While I don’t think this necessarily means a sustained era for na’vi at the very least I purport that the consistency of B1t, in comparison to Flamie, will help elevate the floor of Na’vi that we will see over the next few months.
References:
Parsing framework + Data Model: https://github.com/pnxenopoulos/csgo
Heatmap Visualization: https://github.com/markus-wa/demoinfocs-golang
