We assess how easily each starting kit supports a readable, forgiving run. Every partner has an individual explanation, with community voting alongside the selected card. These recommendations come from current client data, not results from a controlled playtest, a speedrun leaderboard or a measured win-rate study.
Begin with a partner you can actually recruit, then inspect why its strengths fit the problem you are trying to solve. A high rating is useful only when its underlying assumptions match your route, upgrades and comfort with movement. The starter comparison is the better entry point when your choice is still limited to the opening partners.
What the editorial tiers measure
The rating favors benefits that are available early, easy to understand and relevant while learning. Reliable regeneration, mitigation, shields, movement and uncomplicated attack investment receive more weight than a spectacular conditional ceiling. A partner that only shines while injured, airborne or repeatedly applying a particular condition asks for more execution or preparation. That does not make it weak. It means the player has more requirements to satisfy before the plan becomes dependable.
S marks especially approachable combinations of starting utility. A marks a clear, useful plan with a manageable limitation. B marks a more conditional or specialized plan. C marks an opening that asks for more adaptation or a delayed payoff. D remains a valid community choice, but we do not force an editorial partner into it merely to fill a row. The example loadouts translate the same reasoning into concrete moves, items and vitamin decisions.
Start with the reason, not the letter
The short explanation on each card is the important part of the ranking. For Bulbasaur, healing and regeneration explain the recommendation; the letter alone does not tell you that mixed attack classes can complicate investment. For Squirtle, armor and shield provide a buffer, but they do not make standing inside a crowd safe. Abra has explicit protection against ground hazards and certain incoming attacks, yet that still leaves other threats to manage.
Read a promising card, open its listed kit, and ask what would stop the passive from helping you. Lifesteal needs relevant hits. A full-health benefit needs you to recover between mistakes. A contact trigger first requires contact. When the answer exposes a weakness you already struggle with, a lower-rated alternative with simpler recovery may be the better personal choice. Use the build editor to inspect that alternative with the upgrades you actually intend to take.
Recruitment difficulty is a separate decision
An accessible kit and an accessible unlock are different properties. The grid can recommend a legendary partner for its readable combat utility without implying that a fresh profile can select it. Before committing to a new main, inspect its League challenges and distinguish a requirement that accumulates across play from a requirement that must happen within a run. That distinction changes whether the next sensible action is general progression or a deliberate challenge attempt.
The recruitment checklist records the actual challenge text. The legendary recruitment guide explains why the birds, Mewtwo and Mew form different progression targets. Neither rarity nor a famous species name earns a tier automatically. We assess the listed kit and passive; we then let you decide whether the time needed to obtain that partner makes sense for your current goals. Unlocking more options is progress even when your familiar starter remains the most reliable choice.
A starting kit is not a finished build
A partner defines a starting direction, not every upgrade that will appear. A Fire-focused kit can make a type item easy to justify, while a kit split between physical and special attacks asks you to decide which moves carry damage and which provide utility. A status move may earn its slot by protecting movement even when a simple damage comparison makes it look unimpressive. Conversely, an attractive passive does little if the attacks that should activate it rarely reach your targets.
This is why the grid does not rank partners by adding public move power values. The extracted runtime model has its own attack patterns, cooldowns and scaling, and actual runs also depend on upgrade rolls, targeting and conditions. The calculation hub separates supported comparisons from unsupported simulations. Use it to check a specific change, then confirm the behavior in the game rather than treating a large number as a complete combat result.
Understand the limits of this snapshot
The current reference is client v0.8.34, build mvlr7rzv. Runtime extraction includes 816 move records and 106 fusion recipes, but coverage is not the same as a complete battle simulator. Dynamic passives, temporary effects, target overlap, evolved forms and actual movement can change the outcome of a run. Our editorial order intentionally avoids claiming exact damage rankings across those interacting systems.
The species reference also contains public Pokémon information that serves a different purpose. Base-game stat totals and evolution entries are useful identifiers, but they are not proof of PokéBonk's runtime evolution schedule. Where the extraction still reports unavailable mechanics, the honest response is to leave the numerical claim open. A future verified mechanic or balance change may justify revising a reason even if the familiar species names stay the same.
Compare partners with a useful experiment
Choose an outcome before changing your partner. If small hits gradually end the run, compare recovery and mitigation. If you cannot reach a safe lane, compare mobility and crowd control. If enemies remain alive long enough to surround you, inspect damage coverage and upgrade focus. Keep the map, difficulty and broad resource budget comparable so that a new partner is not credited for a much stronger permanent-upgrade profile.
Record the failure point and what was happening immediately before it. A missed exit, a purchase that delayed damage, or a move that attacks in an awkward pattern is more useful evidence than “felt bad.” Repeat only enough to understand the pattern, then adjust the plan. If your test revolves around a combined move, check the fusion recipe reference before reserving slots; a fusion that never becomes available cannot justify weakening the entire opening.
How community voting differs
Community votes answer a broader question: how visitors to this site rate a partner for the displayed version. They are not inputs to the editorial reason. Keeping the views separate allows you to see disagreement rather than hiding it inside a single authoritative-looking letter. The selected partner needs 20 votes before its community tier is displayed; a quiet panel means the threshold is unmet or the service is unavailable, not that the partner has a hidden low score.
Votes are attached to anonymous browser identifiers and version cohorts. Browser identifiers can reset, so the system cannot prove that each identifier belongs to a different person. The displayed tier summarizes submitted opinions, not every player of the game. When you vote, consider the scope you actually know and avoid converting hearsay into certainty. A popular recommendation can be worth investigating while still being a poor match for your unlocked maps or preferred controls.
When to depart from the ranking
Use a specialist when the objective calls for that specialty. A status challenge may justify a partner with a focused condition engine. An exploration goal may favor movement even if damage takes longer to develop. A partner whose mechanics you understand can also outperform a theoretically friendlier pick whose attack pattern you have never practiced. Ratings help narrow a decision; they do not replace the reasons for making it.
The source of factual kit and passive claims is the official game client, preserved with the build identifier above. The tier assignments themselves are this site's editorial interpretation. We will change them when better evidence changes the reasoning, rather than claiming a permanent meta. If you want a practical next step now, pick a card whose weakness is manageable, open its plan, and give the run a specific learning objective.
Frequently asked questions
Is S tier a promise of the strongest possible damage?
No. This list prioritizes starting-kit accessibility and survival utility. A specialist can have a higher ceiling under particular conditions, and no tier guarantees a clear. Read the individual reason and compare the conditions your own plan can reliably satisfy.
Why can the editorial and community letters disagree?
They use different evidence and scopes. The editorial letter follows the method described here; the community letter summarizes eligible visitor votes for the version. Disagreement is useful context, not an error to conceal or a reason to overwrite either result.