Business idea & prototype video created by Ryan Pijai
E-sports (aka electronic sports) are competitive video games that are played just like professional sports are. Teams of players compete against each other for millions of dollars at tournaments. In the past few years, e-sports have grown into a multi billion dollar industry.
- League of Legends
- Dota 2
- Street Fighter V
In these games, there are often key moments/tactical situations where mistakes cost teams entire matches. Players and viewers often watch replays of these to learn from them.
But what if instead of passively watching replay videos, you could rewind game situations and actually try playing them differently?
Professionals and amateurs would be able to practice and experiment with their own previous game situations. They would also be able to practice inside of other players’ game situations! Want to see how you compare against the best video game player in the world? Try playing as him/her in the final major skirmish of a championship match that was played in front of hundreds of thousands of live viewers!
The company I am proposing would build “Tactical Situation Rewinders” (TSRs) for each of the major e-sports. These rewinders allow users to take control of any player at any point in a previously played match. The TSRs then simulate what all of the other players in the game would likely have done given anything the new user decides to do.
The video at the top of this post is something I hacked together based on one of my Dota 2 matches, which involved me and 9 other human players: 4 others on my team (green) and 5 on the opposing team (red). I focused on a critical situation where both teams are fighting heavily. After the fight ends, I simulate rewinding the situation by playing the video in reverse back to the beginning of the encounter. I then play from then on in slow motion.
Although in the prototype video I am doing the same actions after rewinding the situation as I did before it, in the real system users would be able to change their actions and experiment with doing anything they want.
The primary use case for TSRs is player training.
Difficulty Easing: Users will be able to control the speed of the game while training. A user may first try playing a difficult scenario at 25% speed. As the user improves, the speed could then be increased to 50% and then to 100%.
Learning From Others: Users will also have the ability to download how other players handled situations. The movements and actions of those players can then be overlaid as ghosted outlines that move concurrently as users play out the same situations. This will allow users to compare their lines of actions with other players that are better than them and to improve much more quickly.
Pro Team Scouting:
With TSRs, we can make it easier for professional teams to find who they need. Looking for a particular type of player to round out your professional team? Set up a contest consisting of 10 difficult scenarios all applicants need to pass first before being considered.
How will this business make money?
This company will form partnerships with the game creators we are creating TSRs for. Some of our revenue will come from deals we make with our partners. Others will come from premium services we offer directly to our users in the form of downloadable training scenarios and personal coaching.
Securing Partnerships: Since we need to integrate with external game systems in order for our services to work, securing partnerships with e-sports game development studios is a do-or-die type of hurdle we need to overcome. In order to do that, we need to prove to each game company that our services will help them bring in more users and increase user engagement.
Technology: The next biggest hurdle is that the technology for this type of service has never been created before. A replay video contains only data about all the actions that actually took place in a match. It does not contain data about what every player would have done if one of the players did something differently from what was done in the replay.
Will it be difficult to create this technology?
Do I think it is impossible?
No, I believe it is a hard problem, but a solvable one.
One approach to creating the technology would be to use machine learning and to train our TSRs on the massive amounts of play data that already exist for each of these games. We could analyze the actions of all players within a given skill bracket and use that to predict what an average player in that skill bracket would do in any given situation. We could also analyze an individual’s entire play history and use that to help determine what play style we should use for that player. In the end, we will have to look at each e-sport separately and figure out what combination of AI and simulation techniques would work best for modeling players in each particular game.