League of Legends streamer and renowned Shen one-trick xPetu dedicated his entire Master’s thesis to challenging the developers’ philosophy on item balancing. He delves into how item win rates function and critically examines the concept of the meta in the game.
Having been a top-ranked player for several years, xPetu’s success is largely attributed to his preferred champion, Shen, known for his remarkable build flexibility. He stands out as one of the world’s elite players using the Ninja, often showcasing unconventional builds that might appear as troll choices to the average player.
For instance, despite most players steering clear of building Rocketbelt on Shen, he boldly incorporates it into his kit.
In essence, this Shen one-trick based his Master’s thesis on advocating for an innovative approach to champion builds, articulating his reasoning in a manner applicable to any champion, and even created an entire app around his findings.
And yes, he successfully graduated.
xPetu explained his concept about inflated win rates for items, a notion many seasoned LoL players are familiar with. A prime example is Mejai’s, which players typically acquire when they’re already ahead in the game, primarily using it to snowball their advantage.
However, he proposed that this mindset should extend to a broader array of items, questioning Riot game designer Phreak’s philosophy on item balance.
“Generally speaking, if you dig in and say, buy his third most common starting item, and his win rate is actually 2% higher, that is almost always fake data,” Phreak mentioned during a discussion about patch note modifications on his YouTube channel.
“Any item that is not his most built item has a default inflated win rate because the only players who are smart enough to identify that this item is good are the ones buying it. They’re already inflating the win rate by being proficient with the champion.”
xPetu commended Phreak’s insights but countered that the Riot employee’s reasoning was flawed, indicating that it represents “human intuition” rather than a strictly data-driven analysis. Although Phreak’s ultimate conclusions may be sound, the method of arriving at them could be misguided.
Consequently, xPetu endeavored to produce something more substantive.
To illustrate his findings, here is a sample from the data in his 38-page thesis focusing on Shen’s win rates, specifically against magical top laners concerning their first item choices.
Hollow Radiance is a popular item due to its magical protections, evident from its higher sample size (K) in these matchups. The data reveals that while players frequently purchase Hollow Radiance as their first item on Shen in magical scenarios, doing so diminishes its actual win rate (51.25%) in comparison to its estimated impact (52.67% win rate) as per his algorithm.
This summary simplifies xPetu’s thesis and its complexities, yet it effectively conveys the primary idea. While he acknowledges that Riot probably possesses more extensive data than his limited API access could provide, he believes there are meaningful conclusions to derive from his findings, despite lacking comprehensive real-time data.
For example, items often get balanced based on their scaling with a player’s level. With this fresh data, developers could enhance Hollow Radiance’s passive scaling at lower levels while maintaining its same effectiveness later on, making it more advantageous against magical champions early in the game.
Such adjustments could help the developers make informed decisions based on solid statistics instead of relying on intuition or loosely connected data. More precise data would greatly benefit the development team, particularly when addressing balance issues with champions like K’Sante, signifying the value of efficient developer time.
Moreover, xPetu critiqued the very notion of a meta in his thesis.
“Esports communities frequently share popular strategies, which can lead to over-reliance or misapplication of certain widely recommended tactics, causing associated actions to appear worse than they actually are,” he argued.
This highlights the challenge of players assuming they know the optimal approach when, in reality, the strategy they’re employing may be considerably less effective than available alternatives.
This exploration primarily aims to determine the optimal item for Shen. However, xPetu’s insights could potentially transform how players understand the meta and influence how professional players compile their best builds.
Leave a Reply