Gươm Vo Danh and the On-Hit Item Group in League of Legends: An Unverified Balance Signal and the Trap of Early Belief

INTRODUCTION: ONE COMMENT, FOUR LANGUAGES, AND NOT A SINGLE METRIC In Jakarta I usually close my working day with two screens. One shows the data sheet of a Liga 1 football club: high-intensity running distance, duels contested in the final third, expected goals per shot. The other shows League of Legends, and that particular night it showed a Reddit thread I read while waiting for a data export to finish. My phone buzzed a little before eleven. A former colleague from a League of Legends analytics department sent me a screenshot. In it was an exchange involving a designer at Riot Games, saying roughly that an item within the on-hit group needed improvement, that the intention was to improve it for both melee and ranged champions, and that ranged champions should be prioritised. Within six hours that line had been translated into at least four languages. Within twenty-four hours, thousands of people had firm conclusions about a patch that does not exist. Within forty-eight hours, someone had already built a comparison table of champions whose values had not been updated. Across that entire debate, the number of properly sourced metrics I could count was zero. No win rate. No pick rate. No gold-efficiency measurement. No dataset large enough to justify the claim that the item group is weak. All the community had was a shared feeling — that building toward on-hit has become pointless — plus an informal remark from someone with professional authority. That was the entire raw material. And from it, people cooked a feast of assertions. I am not writing this to dismiss that feeling. Community sentiment is often an early signal, and in seventeen years of watching sport I have learned that crowds sometimes see what the data tables have not yet recorded. What I want to do is separate two things that have been blended together: a balance signal, and a belief that formed before the signal was ever verified. Numbers never lie — only the way we listen to them is wrong. CONTEXT: WHICH ITEM GROUP, WHICH PATCH, AND WHO IS SPEAKING Before the analysis, a few foundational facts need pinning down, because history shows that most esports arguments start with two people debating two different versions of the same event. First, the subject is League of Legends. The item at the centre of the argument is rendered as Gươm Vô Danh in the Vietnamese client. One thing few articles are willing to say plainly: Vietnamese client names do not map one-to-one onto original names, and reverse-mapping from the Vietnamese name to the international name is not always exact. The other items named in the discussion are Móc Diệt Thủy Quái, Đao Tím, Cung Chạng Vạng, and an item whose Vietnamese name evokes death's dance. I will use the Vietnamese names throughout and flag clearly where the mapping remains open, rather than assigning an English name for the sake of appearing rigorous. Second, this item group shares a mechanical trait: its effects trigger on each basic attack, not on a single ability cast. This is the crux. An on-hit item does not interact with its user in an additive way; it interacts multiplicatively with attack speed, with attack-reset mechanics, with stacking effects, and with the entire damage structure of the champion using it. Third, the most frequently cited marker in the debate is Patch 25.14. According to what circulated, in that patch Riot Games buffed a series of items in this group — Gươm Vô Danh among them — in an attempt to revive items judged to be weak. That effort did not resolve the problem, or it did resolve it but the effect was eroded by meta drift. Either way, the observable outcome is that the community still believes the group is in a bad state. Fourth, and most important for anyone whose job is reading data: the central statement of this whole story is not an official announcement. It is a community exchange, shared through Reddit, by a designer with a strong public voice. Between a remark like that and an official patch note lies not a few days but an entire process: internal measurement, playtesting, impact assessment across the affected champion class, timing considerations, and finally a decision on whether to ship. In other words, what we hold is a signal, not an event. This is where I want to tell a story from my own work. In March 2026, when I was an assistant analyst at a club in Jakarta, I submitted a forty-page report proposing a role change for a young midfielder. It rested on one very specific finding: he ran only 8.2 kilometres per match, below the midfield average, yet produced eleven passes into the final third — the highest in the squad. I concluded he should play as a number ten, not wide. The coaching staff dismissed it at first, then agreed to trial it after three persuasive matches. The result was two goals and three assists in the next three games. I tell this to make a precise point about my craft. What I presented in 2026 was a prediction model, and it was right. But if it had been wrong, the damage would have stayed inside a closed meeting room, among five adults, before anyone outside knew it existed. What is happening with the on-hit debate is the inverse: the conclusion was published first, the model was never built. And the cost of a prematurely published model is not whether it is right or wrong. The cost is that when the real data finally arrives, people are too tired to read it. BODY: WHY A WEAK ITEM GROUP IS SO HARD TO CURE An on-hit group is a set of mechanics, not a single item The most common mistake in balance discussion is to rank an entire item group because one item within it is weak. What the community debate actually says, if read carefully, is not "Gươm Vô Danh is weak." It is "the whole on-hit group has a problem." The gap between those two statements is far wider than it looks. A single item can be fixed by adding a few stat points. A group cannot, because the cause of a whole group's decline may lie somewhere else entirely: matches shortening, eroding the value of ramp-up phases; burst damage being pushed upward, denying a playstyle that needs continuous attacking time; or the enabler group — items and abilities that boost attack speed and on-hit effects — being adjusted down and dragging the rest with it. Remember this: when a set of mechanics loses value, buffing one item in that set is like giving painkillers to a patient who is bleeding out. The pain eases. The cause does not. The amplification threshold — the problem numbers cannot explain In the exchange that circulated, one idea matters more than the buff itself. The speaker said the item has very strong interaction with certain champions who have special mechanics tied to basic attacks, and that even a small stat change can significantly change those champions' power. This describes what I call the stat amplification threshold. Here is the mechanism. For an ordinary champion, basic-attack damage is a fairly flat function of attack damage and attack speed. For a champion with attack resets, or stacking on-hit effects, or converted auto-attack damage, that relationship becomes non-linear. Each added stat point does not add a fixed amount of damage — it is multiplied by the number of attacks landed within the same window, then multiplied again through stacking effects. The consequence: a change the developer calls "small" can become "large" for three or four specific champions, while remaining invisible for the rest of the roster. There is no neutral adjustment. Every adjustment is a redistribution of power — gained somewhere, lost somewhere else. The community names three champions at the centre: Irelia, Yasuo and Yone. All three build their power directly on basic attacks, with resets or stacking tied to attack rhythm. And all three share a demographic trait: extremely loyal player bases, and very high one-trick rates. That demography is usually ignored in balance talk, but it matters. A champion with many one-tricks generates more frequent update pressure on the item system it depends on, because that community has an incentive to monitor and respond every time a stat shifts. Those three names are not the cause of the problem. They are the loudest voices describing it. Here I want to quote a line I keep pinned in my office: a player's value is not in the contract; it is in every off-ball movement. In football, people pay for goals, but goals are created in places nobody films. League of Legends is the same. People talk about the damage shown on the scoreboard, but real power lives in attack rhythm, in ability ordering, in whether the third auto-attack lands at the right instant. On-hit items are the contract the developer signs with that invisible part. And when the invisible part shifts, the scoreboard does not tell you why. Melee and ranged share one item There is one technical detail I consider the core of the entire story, and it drowned beneath the wave of comments: the stated intention was to improve the item for both melee and ranged champions, with ranged prioritised. Stop there. One item serving two fundamentally different damage profiles. Melee and ranged champions approach basic attacks in completely different ways. A ranged champion attacks from safety; the number of attacks in a fight depends on whether they get closed down, and the value lies in maintaining continuous pressure. A melee champion attacks at dangerous range; the number of attacks is limited by the fight window, and the value lies in converting within a short span. When both profiles live inside one item, any adjustment must satisfy two different optimisation problems simultaneously. That is a structural constraint, not a design choice. It explains why the on-hit group is among the most unstable zones in the game's balance history: it is continually pulled in two directions by two user groups that cannot be served by the same number. Prioritising ranged, if implemented, is a directional clue. It suggests the intended beneficiary is the on-hit marksman archetype — a group competing directly with crit-path marksmen. In that race, whichever side has the steeper mid-game power curve wins, and most public data currently shows the crit path ahead at most ranks. Here I must state my own limits. I have a measurement framework I use when working with teams: the On-Hit Efficiency Index. It aggregates four variables — attacks executed per fight, attack hit rate, damage per attack, and time spent in a continuous attacking state. That framework is useful when I have raw match data. In this specific case, I do not. I have no access to the publisher's internal database, and I have no intention of pretending otherwise. That means every conclusion in this piece must be read with a question mark attached. My model is only bad when I am too cowardly to ask it the hardest question. The 25.14 precedent and the buff-then-revert pattern Patch 25.14 is evidence of one thing: the developer tried to revive this item group, and the result was not enough to change community perception. Three explanations exist, and I rank them by how much I believe them. First, and most plausible to me: the 25.14 buff was arithmetically sufficient but insufficient in match tempo. The item became stronger, but matches shortened to the point where the window a user needs to reach that item's power threshold often no longer exists. A stronger item that arrives after the match has ended is still a weak item. Second: the buff was neutralised by other changes shipped at the same time. This is common in multiplayer balance, and it resembles a problem I once hit in football analysis — improving one variable while another quietly pulls the result back. In 2026, analysing a major tournament from Jakarta, I found that a top national team's pressing index had fallen nearly a quarter against four years earlier, and that was why it collapsed early. Nobody improves pressing by adding running distance. Pressing is a system, not a number. Third: the 25.14 buff did work, and community perception is a recognition lag. I do not dismiss this. It is less likely than the other two, but it exists, and anyone claiming certainty that it does not is selling you a belief, not a measurement. The pattern all three explanations point to is buff-then-revert. It is a cyclical oscillation in this game's balance history: a group is weak, gets buffed, becomes over-strong in a narrow set of champions, gets reverted, and the cycle repeats. Each loop costs weeks to months, and what is lost across those loops is not item power — it is player belief that change is meaningful. I once worked with a coach whose view I still reuse. He told me a good coach treats a defeat as an update, not a verdict. From that angle, an insufficient patch is not a failure. It is a data point — provided the decision-maker reads it as data rather than as confirmation that the old approach was right. Build diversity: the least-discussed and most important metric The real problem with a weak item group is not win rate. It is the number of options removed from the table. When the on-hit group is weak, an Irelia player does not lose the ability to win. They lose the ability to choose. They are pushed toward a single build path, and from a player who could adapt to match conditions, they become a player executing a formula. That is a loss of competitive quality, and it appears on no statistics page. This is why I always track something I call build-path breadth: the number of build directions achieving non-negative win rates against the most common path at each match phase. When it falls, you know a balance zone is shrinking — even if the overall win rate of the relevant champions looks acceptable. For the on-hit group, I believe build-path breadth has been shrinking for a long time, and that is the true cause of the feeling the community is expressing. When people call a group trash, most of the time they mean it no longer offers choices, not that it no longer offers power. That distinction changes the cure. If the problem is power, you add stats. If the problem is choice, you must reopen a branch of possibility — usually by adjusting support items, enablers, or match tempo, not by putting a new number on the item itself. Another story from my work, in a very different setting. In 2026, when leagues paused and matches returned without crowds, I built a report on how the absence of spectators affects performance, and proposed raising high-intensity running distance by about one eighth to compensate for the lost home advantage. My club opened the season with eight unbeaten matches. The staff called me a mad professor, in a good way. But what actually produced that run was not the one-eighth figure. It was the fact that the entire training system, selection method and playing-time distribution were adjusted around the same assumption. The number was only the excuse to reorganise the system. Had I delivered the number and changed nothing else, we would have lost more. That is exactly my concern with the on-hit debate: people are asking for a number, while the problem sits in the structure. THE CONTRARIAN ANGLE: A WEAK ITEM GROUP IS NOT ALWAYS A BROKEN ONE Here I part ways with the crowd. The implicit assumption of the whole debate is: the on-hit group is weak, therefore it needs buffing. That sounds rigorous, but it contains a leap I do not accept: it assumes the relationship between cause and effect is direct causation, whereas what we actually hold is a correlation. The correlation is this: players feel the on-hit group is weak, and the on-hit group is being used less. These travel together. But at least four other causes could produce the same observed outcome without any assumption that the item itself is broken. First: the damage race has a new leader. If burst damage rises relative to sustained damage, a build path requiring three to four seconds of standing and attacking loses value automatically, even with every stat unchanged. The item did not weaken. Its environment became harsher. Second: survivability of the champions using it. The on-hit group depends on continuous attacking time. If the relevant champions' survivability in fights falls due to changes elsewhere, the item's value falls too, even if its numbers are untouched. Third: herd psychology. Once a build path is labelled weak, its pick rate falls. As pick rate falls, its win-rate data becomes noisier, because the sample is smaller and dominated by diehards. A self-reinforcing loop forms: fewer picking it, worse data, even fewer picking it. The group dies socially before it dies statistically. Fourth, and in my view the most important yet least discussed: champion drift. If the champions designed to use on-hit items are tuned down for unrelated reasons, the group will be perceived as weak even though it never changed. Players cannot distinguish a weak item from a weak champion. They only see the result on screen. All four causes produce the same symptom. And if the true cause lies in the last three, buffing Gươm Vô Danh would not just fail to cure the illness — it could worsen it, by pushing power into the champions with the highest amplification capacity without freeing choices for anyone else. That is the biggest tactical blind spot in this whole story. Nobody is asking the right question: if this group is buffed hard enough to become a top choice again, what happens next to Irelia, Yasuo and Yone? The answer, based on the buff-then-revert pattern observed repeatedly, is that they become priority picks beyond tolerance. Pick rate rises. Win rate rises. The community starts complaining about them instead of the items. The next patch tunes them down, while the item group reverts to its previous state. The loop closes, and players lose another cycle of belief. That loop is not the developer's fault. It is a property of the system when you try to balance a non-linear mechanic group with a linear tool: a number on an item. AND HERE IS WHAT I WILL TRACK Four signals, ranked by likelihood of appearing. First, the scope of any change. If the official notes touch only one item in the group, this story returns in a few months with an identical ending. If they touch three or four simultaneously and adjust the enablers behind them, that indicates the diagnosis was correct: a class-level problem. Second, build-path breadth for dependent champions. I will use Irelia, Yasuo and Yone as the observation set, because their player populations are large enough for the data to be statistically meaningful. What I look for is not a higher win rate, but the reappearance of multiple build paths with acceptable outcomes. Third, pick rate by rank bracket. If pick rate rises at low ranks but not at high ranks, the group became stronger without becoming harder — meaning it is strong because of its ceiling, not because of skill. For an on-hit group, that is the worst case, since it implies power has been decoupled from skill expression. Fourth, the release cycle. This game is updated frequently, which means this topic has a short news half-life. If nothing ships within the next two update cycles, the probability of a near-term change should be read as low — and most of the current debate should be archived as a document about community psychology, not a balance forecast. CONCLUSION In my profession there is a mistake worse than predicting wrongly: predicting rightly for the wrong reason, then believing you understood the system. That produces a garbage model dressed in a success, and it will collapse the next time you use it. The on-hit debate risks becoming exactly that mistake, at community scale. If the developer buffs Gươm Vô Danh and the group returns to strength, many will declare they were right all along. Very few will check what actually changed — whether it was the item's stats, or match tempo, or some adjustment elsewhere nobody noticed. For me, the question worth chasing is not how many stat points Gươm Vô Danh will receive. It is whether an item group built on non-linear mechanics can be cured with a linear tool — and if the answer is no, then the entire way we discuss game balance is standing in the wrong place. The next season will answer. Meanwhile I will keep the data sheet open, keep the question mark, and keep a professional habit that has followed me from Jakarta to now: whenever too many people are equally certain about something that has not happened yet, that is precisely the moment to open the spreadsheet and start counting again.