Chapter 460: A 1.5 Million Dollar Bounty!
In the early days of artificial intelligence entering the gaming field, things didn’t go smoothly at first.
Before developing Origin, this company had created another AI intended to break into the RTS genre, only to get thoroughly crushed by human players.
Of course, choosing RTS as the main focus was largely due to the nature of the game genre.
Among the games on the market at the time, only in RTS games could humans and AI fight on somewhat equal footing.
In FPS games, no matter how hard humans tried, they could never beat cheats; and in casual puzzle games, humans had no chance of winning either.
Those puzzle levels that seemed incredibly difficult were only hard for humans; for AI, they were completely effortless.
After all, with so many puzzle games in the world, which one could surpass the difficulty of Go?
As for the currently popular MOBA games, since they were 5v5, having five human players face off against five AIs with identical thought processes? That didn’t seem quite right either.
So, after much deliberation, this AI research company found that RTS games were the most suitable choice.
The key reason RTS was harder to conquer than Go lay in the fog of war.
In RTS, due to limited vision, the AI couldn’t keep track of all the opponent’s movements, leaving room for strategic gameplay between the two sides.
At the same time, because of the extremely complex counter-relationships between different unit types, changes in RTS often happened faster and were more intricate than in Go. If you failed to notice or understand a small intention of your opponent in time, you might get steamrolled in a single wave.
For AI, the biggest limitation was always one thing: computational speed.
In the early days of AI versus Go, humans could beat AI in fast-paced games, but once the pace slowed down, they would lose disastrously. This was because AI and humans thought in completely different ways—the more complex the situation, the more time the AI needed to calculate the optimal solution.
In real-time strategy games, unlike Go, there wasn’t ample time for thinking. For AI, this might require a completely different algorithm and set of rules.
Of course, aside from these aspects, AI completely dominated human players in terms of micro-management.
For human players, APM in RTS games was never enough, because you needed to control dozens of units simultaneously. Even if your brain could keep up, your hands couldn’t.
But for AI, controlling any number of units was no problem, and whether in reaction speed or accuracy of operations, it completely crushed human players.
So, considering the relative unfairness to human players, the company behind Origin imposed certain restrictions on the AI’s micro-management, forcibly capping its average APM at around 400 and its peak APM within 600, matching the level of top-tier human players.
Of course, even with these APM restrictions, the AI still had an advantage over humans because it didn’t need to perform repetitive actions to warm up—every single one of its actions was effective.
In other words, while the average appeared to be around 400, it was actually equivalent to a human’s effective APM.
After Origin successfully topped the ladders of the American, European, and Korean servers, its development company announced that, considering Origin’s excessive advantage in APM against human players, they would further lower the cap to an average of around 350.
Of course, this could be seen as a move for fairness, or it could be that the company feared scaring away professional players, leaving no one to challenge, or worried that the man-versus-machine match would become one-sided and lack entertainment value. Hence, they made this adjustment.
But for Warcraft professional players, this felt like a blatant provocation.
At the same time, Origin’s development company announced a bounty of 1.5 million US dollars for this man-versus-machine match—anyone who could defeat Origin in a best-of-three would claim the prize!
This massive bounty was a huge temptation for Warcraft professional players worldwide. After all, Warcraft wasn’t as hot as League of Legends, and these players weren’t exactly rolling in cash.
Moreover, this wasn’t just about the money—it was a chance to gain fame! Fighting on behalf of humanity and defeating AI was a supreme honor that most people would only encounter once in a lifetime. Who wouldn’t want that?
Even if Origin’s company hadn’t offered a bounty, these professional players would definitely have stepped up.
And Starlight, the “Number One Elf” on the Korean server, clearly held this determination, becoming the first professional player to face Origin head-on.
...
Online, some people wondered why, since Chen Mo had developed Warcraft and was widely recognized as the top Warcraft player on the Chinese server, Origin didn’t challenge him directly, instead going after Starlight, the Korean server’s top Elf player.
Actually, this wasn’t hard to understand.
In Chen Mo’s previous life, AlphaGo didn’t challenge Ke Jie right away either; it started by taking on ordinary top-tier players step by step.
On one hand, this was to build up buzz gradually. Challenging a world champion right off the bat wouldn’t generate as much hype or attention.
On the other hand, it was about taking things slowly. Since Origin’s company didn’t know exactly how strong Origin was, climbing the ladder step by step gave it a better chance of winning.
Moreover, choosing Starlight was indeed the best option. They had just played on the Korean server’s ladder, and the gap between them wasn’t huge. Plus, both players used the Elf race. While mirror matches might be slightly less exciting to watch, they guaranteed absolute fairness.
If they had chosen different races, even if Origin won, some might claim it was due to racial advantages. Starting with a mirror match better showcased the fairness of this man-versus-machine contest.
Even before it began, this opening match of the man-versus-machine battle had already captured the world’s attention.
It was called the opening match because everyone knew that, win or lose, Origin wouldn’t stop there. It would only grow stronger and continue challenging even tougher human players.
Opinions on the match’s outcome were divided.
“A week is enough for Origin to improve by leaps and bounds. Starlight couldn’t beat it before, and now he definitely can’t. We’ve seen this script countless times in previous AI-versus-human board game matches—what’s there to doubt?”
“But this time, they’ve further lowered the APM cap. RTS games are all about interconnected systems. If Origin’s hero gets killed due to insufficient micro-management, wouldn’t the whole game be over in an instant?”
“But offline matches have no lag. Last time Origin struggled mainly because it was playing on the Korean server from North America with latency!”
“Last time, Starlight wasn’t prepared. If he studies Origin’s replays carefully, he’ll definitely come up with counter-strategies!”
Players were all talking at once, each holding their own opinion.
Although most players were rooting for humanity, they also knew that Origin had a higher chance of winning.
After all, this kind of script had played out many times before in board game battles.