Before I continue with my Q-Analysis, I wanted to share some thoughts and data I collect on a certain topic from APEX Season 2 tournament.
This post is all about a concept, “First Blood”, that has been around in other e-sports and online games, but not so much in Overwatch because of the nature of the game. However, after doing my analysis on the Control Controversy and after noticing how different of a gamemode Control is, I thought to check how this first kill aspect fares in Overwatch. If it has any kind of impact, what kind of impact and are there any conclusions on can come to just by analysing first kills/first deaths?
You will be surprised by the results in the context they are going to be presented.
For better viewing experience, open images in a new tab to see them magnified.
The first kill
Before I begin my analysis about this specific concept in Overwatch, I want to make a brief comparison to other e-sports, like CS:GO or LoL. This will also help everyone understand a bit my parameters and assumptions on how I will approach it from an analytical point of view.
In CS:GO the opening kill is a big deal, because death in that game is permanent till the round concludes and a new one starts. Therefore, a team losing a player first is actually putting that team at a disadvantage for future engagements, numbers-wise at least. In a game where teams are splitting, like they do in CS:GO, this one kill can basically shut down a specific play or tactic, can easily foretell how the whole round will unfold to the eyes of casters and viewers and at the same time -most importantly- is putting the pressure on a specific team to clutch, raising the tension and creating the atmosphere for big plays that everyone wants to see. Still, all first kills are not the same and a lot of parameters must be taken into account to properly evaluate their impact.
In LoL, where First Blood is announced explicitly in-game with text and audio, this opening kill holds significance in terms of outplaying your opponent in either a duel, a skirmish or a big engagement. It depends a lot on the players and their style, the team’s plan for the first fight and so on, but since there is much more room for individual duels and plays at the beginning, the First Blood is actually an opening act that star players do more often than not. If someone else draws First Blood, it also tells volume about this player’s skill and performance.
In Overwatch, however, we witness a total different approach in this term. All engagements are team engagements, especially the first engagements. It’s really unlikely that the first kill will come as the result of a duel that is happening far away from the bigger engagement. Overwatch is a heavily team-focused game and with all the damage dealt per fight not actually conveyed by any means, the first kill can not really be hyped that much. It can also be something totally random: a lucky strike that a player did and caught an opponent at low health and finished him off. Nonetheless, in order to actually prove the significance or not of such an opening act in Overwatch, research must be done, which I have started working on and I will present straight away.
[All data can be seen in the end of the post in a cumulative table with all stats]
Analyzing such a specific moment in the whole game can really provide little to no interesting facts, if it is done in the wrong way.
My parameters for this analysis are the following:
– I have examined only the first round of matches from APEX S2, where all teams have played at least once against another team.
– First Blood is divided on a per map & per round of a map and per map & per side of attack/defence basis for the different gamemodes.
– I have written down who drew First Blood, who was the first death in the round, the time that it happened, which team and what hero each player was playing.
– I also have written down the winner of the first engagement, after the first kill and, for Control maps, the winner of each round.
[Winstonslab.com timeline really helps with tracking all the necessary data.]
In short, I have tried to see how First Blood is actually affecting the first engagement in general, which I will explain in the next paragraph why I think it is important, and -for Control maps again- how this first engagement has affected the result on the round. The other data is actually the data one would expect to see in such an analysis: what heroes are getting more first picks, what heroes are getting picked first, the heroes’ role in the team etc.
The first engagement
Overwatch is a game with really many levels of strategy that are actually affecting each other in lots of indirect ways and I imagine that teams have yet to fully master the way to engage in every occasion, or how to make the most out of each round like the teams in CS:GO have achieved, for example.
One of these levels of strategy concerns the first engagement and the result of it, because it affects the Ultimate Economy in a really big way. In Control the impact is much bigger than other gamemodes (Escort, Assault, Hybrid), because the winner of the first engagement actually gets first turn in defending the objective, gets more Ultimate Charge than his opponent and can up to a point set up for the next engagement in a gamemode with no organised attack/defence sides.
Also, losing a player up front can really decide the outcome of the match depending on the way that first kill was delivered, the role in the composition and the opponent’s composition. Because, as we will see in the data collected in the next paragraphs, a quick loss does not really guarantee victory in the first engagement. And the same goes for the first engagement. Losing it does not guarantee that the team is gonna lose the round on a Control map, the only gamemode that is so heavily affected by the outcome of the engagements. This is the only gamemode with no pre-defined timebank and depends on the outcome of the engagements how long each round will last.
First engagements, however, differ in a lot of aspects. For this analysis, I have defined first engagement as the engagement that begins with the first death and either leads to a team wipe or a retreat. Even if only one player remains on the battlefield and fights against the opponents after the first death on either of the two teams, the engagement still goes on. To consider the engagement finished, some seconds must pass where players are not actively engaging. [Analysis about the length of engagements after first blood will be done in an another post.]
So, let’s analyze the data for first blood and first engagements on Control maps.
First blood on the Hill
So far in the first round of matches in APEX S2 Group Phase (where all teams have played one match), a total of 9 Control maps were played (1 was an Overtime Bo1) with a total of 20 rounds.
During those 20 rounds in 10 ocassions the team that drew first blood also won the round.
Only 9 teams out of the 16 also managed to draw first blood and win the round.
LW Blue was the only team to manage it in two rounds.
So, what this basically means, is that in 50% of the rounds played in the Control gamemode, the team that will draw first blood will probably also win the round.
What’s more interesting, however, is the connection between first blood, first engagement and the win of the round.
As you will see in the following table, only 3/20 times the team that had the first death managed to win the first engagement. In all other occasions, the team that managed first blood also secured the first engagement.
Additionally, in all 3 occurrences of a team suffering first death but winning the first engagement, this team secured the entire round.
The teams achieving this are the following:
Out of the 17 times that teams had the first death and lost the first engagement, only in 6 cases teams managed to win the round. Their names can be seen in the table below.
Finally, the times teams managed to succeed in all three aspects (first blood, first engagement and round win) were 10, 50% of the rounds.
The next level obviously is to see which Heroes/Roles in-game were the first death in this specific gamemode.
Data shows that the most first deaths for Control maps holds the Tank division and more analytically:
Reinhardt (4 times)
Zarya (3 times)
Winston (3 times)
D.Va (1 time)
Next in line is the DPS division with 5 first deaths:
Soldier76 (2 times)
Genji (1 time)
Tracer (2 times)
Final in line is the Support division with only 3 first deaths:
Ana (2 times)
Mercy (1 time)
So far, no Lucio has been the first blood on a Control map and seems like D.Va has also increased survivalability in the first fight compared to Reinhardt, Winston and Zarya. Roadhog is also a hero that is missing from this list in the Tank division, if we take into account that he’s almost a standard pick pre-patch and post-patch.
First blood on Attack/Defence gamemodes
For the rest of the gamemodes (Escort, Hybrid, Assault), I have done my analysis on an attack/defence basis per gamemode.
As you will see in the table below, for Hybrid modes the ratio for first blood on attack is the same as on defence (50%).
For Assault maps, however, the First Blood ratio on the Attack side is overwhelmingly bigger than the one for the defence side (61% vs 39%). What this essentially means is that on these specific maps, the attackers have much bigger probability of getting the first pick.
For the Escort maps, the ratio for first blood is in favor of the defenders (40% for attackers vs 60% for the defenders). This means that defenders have shown a higher possibility to get the first pick in the first engagement.
The gamemodes are not seeing equal representation in this tournament. I have highlighted this in the table, where you can see that escort maps have a much lower appearance rate than Assault and Hybrid for the simple reason that they are the fourth map in rotation and matches can be decided before teams reach them. So, always keep this in mind.
First Blood on both Rounds
In these gamemodes the interesting facts are revolving, of course, around the number of times teams managed to draw first blood on both sides of the map. The result is a bit more than 50%. 9 out of the 16 teams managed to get the first kills in both sides of the map and you can see their names in the following list. Only 3 teams, though, managed to draw first blood on both sides on two different maps: MVP Infinity, EnvyUS & LW Blue. Moreover, the total amount of maps that teams managed to secure both first bloods were 12/21, more than the 50% of the maps played.
Still, the number of teams that achieved this feat and managed to fully take the map was much less: 5/16.
Let’s go even deeper in the data.
The maps that teams managed to secure both the first kills were the following:
King’s Row (2 times)
Temple of Anubis (2 times)
Hollywood (2 times)
Volskaya Industries (1 time)
Hanamura (1 time)
Eichenwalde (1 time)
Watchpoint: Gibralter (3 times)
In the table you can see also the number of times the maps were played in these specific matches.
Fun fact from this data is that Watchpoint: Gibraltar is the map on which teams drew first blood on both sides most times. It was played 3 times and teams managed to secure both first kills all 3 times.
The gamemode with the most double first bloods is Hybrid, however, which is also the most played gamemode, for the reason we already explained before.
In the next tables you will see the performance of heroes in all the gamemodes and you can see not only how many times they were the first death, but also how many times they were the ones that drew first blood and the overall ratio of the heroes who managed to do both.
One will immediately notice that Reinhardt is reigning all over the First Death list with a stunning 16 deaths as the first pick on a map. The overall rounds played in these matches were 68, so it’s the first death every 4 rounds. At the same time he’s the hero with the most first bloods drawn as well, so his overall chance to draw first blood over dying first is 75%.
Another immediately noticeable fact is about Roadhog, who has drawn first blood in 11 rounds out of the 68, but has been the first death only 3 times in these matches. This means that he has an overall chance to get the first blood than being killed first of 366%. More interestingly, in the pre-January patch his first bloods were 7 over 1 first death, while post-patch they were 4 over 2.
This does not mean by any sense that his overall potential has dropped, because the stats are taken from distinctly different in skill and unique teams. As the tournament progresses, comparing stats from the same players or teams will give us a more accurate view of the Roadhog situation.
In general, Tanks seem to secure the first kill a lot more times than DPS heroes do, which one would claim that it is their role. Still, this does not mean that DPS heroes do not do damage. It just shows that Tank players seem to finish off targets a lot more than the DPS players and this is interesting because when analyzed in the context of the compositions played by each team, it will provide us with more interesting results. [Will be done in a future post concerning this issue.]
It’s not surprising that Tanks are also at the top of dying first in the first engagement, but it comes as a surprise that Supports have much bigger survivalability than DPS heroes in the first engagements. Again, this is very intriguing and needs further investigation, but let’s take the example of a Triple Tank composition. Since it is much more probable to lose a Tank or a DPS player up front, immediately your composition is starting to fall apart, because all three tanks are playing a vital role and most teams are running one DPS with them, which means that their big cannon is being killed 1/3 straight away. More analysis will show us in the end how important DPS heroes are in those scenarios and how the compositions on a per gamemode basis affect these stats.
Of all DPS Heroes, Pharah shows the worst performance for this role with 1 first blood over 2 first deaths. Genji has a ratio of 4/7, with tracer having a 4/5 and Soldier76 topping all of them with the neutral 4/4 (first blood / first death). Up to a point, all this data is explained simply by the health pools of the heroes. DPS heroes have significantly lesser health pool and can be one-shotted with ease.
Fun Fact: Neither McCree or Reaper have drawn first blood yet in any round.
Out of all support heroes, Zenyatta seems to be the one drawing first blood more often and getting killed less. Even Lucio has been killed 2 times right off the bat in those matches, while Zenyatta only once. Still, I need to collect more data to compare the times that Zenyatta was in the starting composition in comparison to Lucio’s almost insta-locked presence.
Finally, with this data we can also evaluate the performance of heroes like Widowmaker, whose job is to draw first blood and signal tο others to go all in, since they have the numerical advantage. The data so far shows that most teams have gotten value out of Widowmaker and the players have indeed secured the first kill a lot (4 times already). On the other hand, she has been the first death of the game 3 times, which really makes one wonder about the player’s skill or be amazed about their opponents’ shut down tactics.
The final part is, as one would imagine, a “name ‘n’ shame” analysis with stats concerning the players that died first in these rounds. (Just kidding)
In this specific part I have done only a “first death analysis” because I wanted to see the trends concerning two specific aspects:
1) Is there any trend for teams with a lot of first deaths and their overall win rate?
2) Are the players with the most first deaths actually a liability to the team in general?
The results can be seen in the following table, which I have structured on a per match basis rather than a generalised one with all players in one big list. The reason for this is because we do not have enough data from APEX S2 matches to actually compare the quality of these players’ first death.
Maybe Cocco was focused too much by another team. Maybe Panker was playing tοo aggressively. These are the players with the most first deaths overall. But, Cocco’s overall performance is betther than Panker’s, per se, because Cocco played 11 rounds while Panker played 10. The decimal number is there to show exactly this.
Also, Kisu’s first two deaths in the Flash Lux vs Runaway match as Mercy is speaking volumes about this player’s positioning when they try to play a Pharmercy combo rather than the 4 deaths of Panker who is a tank with less mobility in the front line and can be focused hard in any situation.
Finally, there is indeed a trend between first deaths and match wins, but this needs to be examined closer when more data will be available. For now we can see that usually the team that has claimed most first bloods will probably emerge victorious in the end.
The fun fact in this category is the match between BK Stars and Meta Athena, in which BK Stars actually managed to secure more first bloods, but Meta Athena won the maps.
Lunatic-Hai vs AF Red is another example of teams having almost same stats, “first death”-wise, but Lunatic-Hai was more dominant overall. Same with Flash Lux vs Runaway. It’s interesting to note that in both those matches, only 3 maps were played.
In whole, in 4 matches only 3 maps were played, but the total rounds are not the same. Only 2 of them had 6 rounds played (Runaway vs Flash Lux & Kongdoo Panthera vs Fnatic), whereas the other two played Overtime Rounds as well on those 3 maps (Meta Athena vs BK Stars and Lunatic-Hai vs AF Red).
Conclusions & Future Analysis
Again, I need to stress out that all this data collected is not to be used for immediate conclusions and performance evaluation for either teams or players. It’s more about to identify some trends that one can keep an eye out for as the tournament progresses and in the end to reach some conclusions.
This analysis provides little to no actual evidence outside of the APEX S2 format.
APEX S2 is the best tournament for this type of analysis because of the way it is structured; the map pool on one hand and the teams’ level on the other. No other tournament features so many great teams that are so focused on bringing their A-game into every match.
Comparisons to previous tournaments played on other patches will also provide few interesting facts, mainly on the heroes first deaths/first bloods because each patch favors specific compositions.
So, in the future I aspire to collect more data with the help of winstonslab.com’s excellent system of collecting and presenting data and to make more in-depth analysis on not only the first engagement but also for all engagemens, in relation to Ultimate worth and team compositions.
Keep an eye out for them.