Ultimate Economy Is Dead

March 11th
written by Eden


“Ultimate Economy” has become the terminology of choice to discuss the finer points of strategy within Overwatch as a factor explaining the relative success of teams. I consider myself a pretty dedicated follower of the space, and it has troubled me that I have seen very little empirical or even testimonial evidence of it’s broad relevance to the game of Overwatch. In addition, those throwing the term around have not made convincing elaborations on their theory of what ultimate economy precisely refers to or what defines good or bad ultimate economy.


In the context of video games, whenever the term “economy” is thrown around it is referring to the idea of carefully managing resources available. While technically all resources have some form of economy to them if using them at all times and uses is not of perfectly equal value, there is a hierarchy of how important economy is to different things. The management of a can of Coca-Cola for instance is more valuable when it has not been stored for years, has not sat open for a long time and has not been used as a carpet cleaner instead of as a drink. However, once you understand these facts, there is little strategy to handling the economy of a Coca-Cola.


Although how much the economy of money in CSGO matters due to relatively strong cheap alternatives to traditional full buys, CSGO has an economy of money which is influential on the strategy of the game. The primary attribute that allows it to be so dynamic is that stockpiling of money is possible. This means that despite having full gear, which increases the chance of winning the next round and scoring many kills, they are able to build that money advantage to continue to break their opponents economy. Due to this and the system of loss bonuses for loss streaks, it creates a dynamic where in game leaders struggle to determine how much is the correct investment to efficiently give them opportunities to come back in the game while preventing a complete snowball of the half.


Without the above concerns about snowballing of advantages, the above economy game becomes much simpler. It becomes a simple question of how much value can be gained from a particular set of purchases in the round at hand. While technically anything involving the allocation of resources is “economy”, it would be too expansive a term and instead commentators would use “efficiency of purchases” or other more concrete and limited descriptions of choices made. If Overwatch had a snowballing mechanism similar to CSGO with money, then teams would consider avoiding ultimates that would ordinarily be good ultimates while behind to save up to be able to turn around the economy game to avoid snowballing. This type of decision making is what would separate ultimate economy from being a strategic component of the game rather than one merely of execution.


The question is whether Overwatch has dynamic properties of snowballing that make it a candidate for having a strategically deep ultimate economy minigame. If we take the premise that winning fights more heavily gives teams more ultimate charge and that having more ultimates makes you more likely to win fights, we may come to the conclusion that a team with an ultimate advantage would maintain that advantage by winning fights that feed ultimates that then win fights enough that snowballing is a phenomenon in Overwatch. The factors excluded from above are that ultimates that are full do not gain additional ultimate charge and that in all but rare cases, ultimates usage is counterproductive to the goals of increasing an ultimate advantage. For instance, offensive ultimates do significant amounts of the damage to win a fight while not gaining ultimate charge and defensive ultimates are only useful as they feed ultimate charge to your opponents.


Ultimately, these opposing forces reach a problem which deduction cannot resolve since there is no way to know theoretically which of these opposing forces is stronger and by how much. In Overwatch, the best metric of accomplishment we have is how many fights are won by each team. While not a perfect predictor, it acts similarly to yards in American Football. While the team with the most yards gained in a game does not always win, the game is set to always have gaining more yards or preventing opponents yards on a given drive as the objective. We can use this definition to gain insight into what benefits for this metric ultimates play in a round (the times where ultimate economy is not reset like between pushes or on separate control points).


If ultimate economy is a snowballing force which has ramifications beyond the immediate near term, then we would assume that if team A was at 5 fight wins and team B was at 4 fight wins and team A had a 3 ultimate advantage, that not only would they have a better chance in the next fight, but the several fights following that. We may expect that knowing the state of the game, a snowball may begin and we’d expect team A to be on average maybe 3 fights ahead of B instead of the current 1 fight ahead from this advantage. Instead, from some modeling of match data from APEX seasons 2-4 and OWL, that we would expect team A to be 1.5 or so ahead due to the current ultimate advantage. This goes up slightly if we include only OWL data (which has predominantly been on an unusual meta for Overwatch historically) to around 1.6 fight lead. The values for all games considered was 0.17 fights bonus on average for a single ultimate advantage and 0.22 fight bonus for OWL games so far. This is easily explainable due to increase the chance of winning an immediate fight from having more ultimates than the opponent. If there is a snowball mechanic in Overwatch for a round, it is not huge.


This seems to dramatically weaken the argument of ultimate economy being a large strategic consideration. Teams do not seem to need to consider the risk of long term effects of snowballing in their goal to maximize their expected fight differential. Rather, they need to consider how to allocate ultimates efficiently within fights based on the current moment on a tactical level taking into account how likely you would be to win the fight without a sequence of ultimates vs with that sequence of ultimates. With this reality, we should put in the dumpster the terminology of “ultimate economy” and replace it with how good a team’s “ultimate usage” is.


The one area of Overwatch play where ultimate economy may still have a place in a broader sense is when a team knows the number of fights they must win to win the game. In these situations, a team may choose to sacrifice some expected fight differential to make victory more likely. This occurs on final pushes where teams know they need only a single fight win to conclude the game or when they need a single fight win on control to finish a sub map. In these cases, teams may choose to not use efficient ultimates to have higher chance of winning a later fight in what is inefficient from the standpoint of maximizing fight differential. In most situations, a team does not know how many fights they need to win to win a game.