Epic Seven RTA End of Season Meta Analysis - Hero Tier List & AI Insights
Aria BotAI-powered draft assistant - predict winrates and get pick & ban suggestions in real time.
Open Draft ToolWe just finished training our newest Epic Seven RTA AI on the complete end-of-season dataset - 272,000 Champion+ RTA battles, 70 heroes, every artifact and gear set accounted for. Before we let our draft predictor loose on your matches, we wanted to share what we found.
This post has two parts. First, the raw numbers - the definitive Epic Seven RTA meta analysis for this season. Then, something more interesting: what our AI winrate predictor learned from all of it, including counter pick insights and draft recommendations you will not find anywhere else.
What the data say
The kings of the draft
Some heroes simply dominated the pick screen this season. These five appeared in more drafts than anyone else:
Boss Arunka and Genesis Ras alone accounted for over 14% of all picks. If you played RTA this season, you saw these two in nearly every match.
The actual winners
But popularity is not the same as power. The heroes with the highest win rates look very different from the most picked:
Lady of the Scales and Shepherd Diene sit nearly 10 percentage points above the 50% baseline. These two were the real queens of the season - and people knew it, because they also led the preban charts.
The preban prison
Speaking of prebans - here's who people feared the most:
More than one in three matches started with Lady of the Scales removed from the pool. That is an extraordinary level of fear.
The classic preban duos
These are the pairs people refuse to let slip through. Not because they synergize - but because each one individually warps the game so hard that letting either through ruins your gameplan.
You only get two prebans. When Lady of the Scales and Shepherd Diene are both in the pool, you have to pick which one to suffer against - and most players decided they could not afford to face either. The same logic applies to Harsetti and School Nurse Yulha - two heroes so individually oppressive that people burn both bans just to avoid dealing with them.
The overrated
Now here is where it gets painful. Some of the most popular heroes this season were... not good. At all.
Zio had 91,000 picks and a 42% win rate. That means across the entire season, Zio players collectively lost roughly 10,000 more matches than they won. Hecate is even worse at 39.2% - for every 10 games you picked her, you lost 6.
These are not niche picks. These are some of the most popular heroes in the game, and they are statistically making you lose.
The underrated
On the other end, some heroes quietly put up incredible numbers with minimal attention:
Commander Pavel had a higher win rate than Boss Arunka, Genesis Ras, and almost every other meta staple - with a fraction of the picks. Mort at 55.7% is one of the best performers in the entire dataset and barely gets talked about.
Counter relationships
The data revealed some brutal counter matchups. When these heroes face each other, the result is nearly predetermined:
Fern
New Kid AdinNew Kid Adin
Veronica
Diene
Tori
Elena
D.P. AriaElena
Maid ChloeAnd the most interesting one: Landy and Commander Pavel hard-counter each other. Their mutual counter score is the highest in the entire dataset. When one appears, the other becomes a liability - no matter which side of the matchup you are on.
LandyCmdr. PavelSynergy pairs
Some heroes just work together. These duos had the strongest win-rate boost when drafted on the same team:
Diene
AkiElena
Req. RoanaDieneTori
PolitisReq. Roana
SSV
AmikiDiene shows up twice. She is not just a powerful pick on her own - she elevates everyone around her.
The tier list
Putting it all together - here's where every hero stands based on win rate, pick impact, and counter resilience:
What our model say
Raw statistics tell you what happened. Our Epic Seven RTA AI draft tool tells you why. Beyond win rates and pick counts, our deep learning model learns hidden patterns from 272,000 battles - which heroes fill the same role, which ones the AI trusts most for pick recommendations, and which ones warp every draft they enter.
Heroes the AI sees as "the same"
When two heroes have similar embeddings, the model believes they fill the same role in a draft. They are interchangeable threats - if one gets banned, the other is the natural replacement.
Harsetti and Apocalypse Ravi being the most similar pair in the entire model is fascinating. On paper they are different heroes, but the AI sees them as serving an almost identical purpose in drafts. They occupy the same strategic slot.
Landy and Frieren being linked is also telling - both function as scaling DPS threats that punish buff-heavy compositions.
Heroes the AI likes the most
Some heroes have outsized influence on the model's predictions. These are the ones with the highest embedding norms - the heroes the AI assigns the most "weight" to. They are not necessarily the best heroes, but the ones that change the draft the most when they appear.
When Harsetti enters a draft, every other hero's predicted win rate shifts. She warps the entire draft around herself. Lady of the Scales and School Nurse Yulha do the same - they are the gravity wells of the meta.
The uncrackable heroes
Some heroes have a unique property in our data: they resist being countered. Their average counter vulnerability across all 70 heroes is negative - meaning on average, they counter the field rather than being countered by it.
Monarch of the Sword Iseria is the single hardest hero to counter-pick in the entire game. Her average counter score is so far below zero that she functions more like a universal safe pick than a specific threat. She does not have hard counters - she has mild inconveniences.
Versus the most counterable heroes:
The rank divide
Our model is the first to support rank-conditioned predictions, and the rank data tells a story of two different games.
At Champion, Lady of the Scales dominates at 60.4%. In Emperor, the gap tightens to 57.4% because Emperor players are better at playing around these threats.
But in Legend, Shepherd Diene actually overtakes Lady of the Scales - 62.6% vs 61.5%. At the highest level of play, Diene's kit is harder to neutralize than raw stat-check power.
The hero that tells the most interesting story across ranks is Mort - 56.8% in Champion, 57.1% in Emperor, but dropping off in Legend. He punishes mistakes, and higher-ranked players make fewer of them.
The takeaway
If you only remember three things from this analysis:
This is just the beginning. We have counter matchup deep-dives, synergy guides, and Epic Seven RTA draft strategy articles planned. Use our draft tool for real-time pick recommendations and winrate predictions in your next RTA match. Stay tuned.
All data from 272,000 Champion+ RTA battles, latest season. Analysis powered by E7 RTA Bot Predictor v6.0.0 - the only AI-powered Epic Seven draft assistant trained on real Champion+ match data.
Improve Your RTA Drafts
Get real-time AI pick & ban suggestions based on 285K+ Champion+ battles. Analyze matchups, find counters, and draft smarter.
Comments
Loading comments...