The ego is a very powerful tool in the tool bag that is being a human being. No one is without ego. If they tell you they don’t have an ego, then it’s time to call Sarah Connor, because the robots are going to take over.
How exactly does ego fit into gaming? Winning or losing consecutive matches in any video game will affect your decision-making. Winning in DOTA is obviously the goal. That goal rings true for any video game. Nobody buys or plays a video game with the intention to lose. It’s as simple as winning making you happy and losing making you unhappy.
Depending on the person, losing in DOTA could lead to making riskier decisions that can either pay off or fail miserably based on random variable outcomes. One of the random variables is, of course, the decision-making of the enemy human player. Their ego will also affect their decision-making depending on if they’re winning or losing the lane.
So, what does all of this mean? Elon Musk-founded research lab, OpenAI, arrived on the DOTA 2 scene in August of 2017 by defeating the world’s best players 1-on-1 using only the mechanical knowledge of the game that it learned, and most importantly, without an ego. Tens of thousands of games generated 180 years’ worth of gameplay in which the bots learned from their mistakes and never made them again.
“Aww c’mon, I’m sure I’ll get it right the next time!” says no AI ever.
How does the human factor of ego come into play here? Imagine that you’re playing a video game where an action that you took that should have cost you the game, miraculously swung in your favor and won the game for you and your team. You and your teammates are jumping for joy and patting you on the back telling you how that big risk paid off and how that wouldn’t work in a million years.
Your ego is now inflated, you’re thinking of taking bigger risks because you want to continue basking in the light of that glory. The next round hits and, again, it’s up to you to make a smart decision or a gutsy call. You try going for the gutsy call because you’re making another decision based on your ego. This time it doesn’t work out so well and your teammates are disappointed in you.
How likely are you to stop making that risky call? It’s a gamble.
These bots do not have that ego getting in their way to make gutsy calls. They are on a reinforced learning system. The computer learns when something good has happened and when something bad has happened. The bots will continue to do what works mechanically without the big risks that could potentially cost them games.
In the end, playing against human characters is fun because of the social interaction and the ever-looming chance they’re going to make riskier decisions based on their ego alone. Against bots, it would be more mechanical. You would need to fight against a machine that clicks faster, thinks faster, and performs faster than you. It’s the real John Henry situation of bots in DOTA 2.