Draft Simulator · Mobile Legends

My laneChoose your lane to start the analysis
Rank
0 / 10Choose your lane
Choose your lane: your top 3 picks show up here
Allies
0/5
  • EXP
    —
    EXP
  • Jungle
    —
    Jungle
  • Mid
    —
    Mid
  • Gold
    —
    Gold
  • Roam
    —
    Roam
Composition0/5

Durability, offense, ability effects and damage mix appear with the first pick.

Enemies
0/5
  • EXP
    —
    EXP
  • Jungle
    —
    Jungle
  • Mid
    —
    Mid
  • Gold
    —
    Gold
  • Roam
    —
    Roam
Composition0/5

Durability, offense, ability effects and damage mix appear with the first pick.

Recommended picks
Recommended picks
for your lane
◆

Choose your lane to start the analysis

Pick a lane above (EXP, Jungle, Mid, Gold, Roam) to see the recommended picks here.

◆

Choose your lane to start the analysis

Pick a lane above (EXP, Jungle, Mid, Gold, Roam) to see your recommended picks here.

Source. Mythic · hero stats: 7-day window, updated · pair stats: 15-day window, updated . Official Moonton data (GMS feed), single source — not cross-checked against a second provider.

How to read the score

This tool does not know how to play Mobile Legends. It reads Moonton’s published win rates and shows you which hero the numbers favour for your lane, given the heroes already locked.

The score is a draft score built around 50. It ranks heroes for this draft; it is not a win rate. Select a recommended pick to see its bars and the figures behind them.

  • Hero stats: the hero’s win rate at the rank you chose, over the last 7 days, read against its lane’s average and tempered when the hero is rarely picked, plus a small share for how often the lane picks it.
  • Vs enemy team: how the hero’s win rate shifts when each locked enemy is on the other team, over the last 15 days.
  • With allies: how it shifts when each locked ally is on the same team, same window.
  • Team needs: a small bonus when the hero brings the damage type or the frontline your locked allies lack, read from Moonton’s official hero profiles.
  • Counter-pick risk: a small plus for a hero with few bad matchups in the enemy lanes still open, a small minus for one that is easy to counter. It shrinks as the enemy locks heroes.

A bar marked not read means no figure exists for this draft yet — for example no enemy is locked. The ± next to the score is the uncertainty of the pair figures behind it.

What is measured, and what is not

Everything shown comes from one source: Moonton’s own statistics feed, the one behind the hero rankings on mobilelegends.com. It is not cross-checked against a second provider, because none publishes the same figures independently.

  • Measured: win, pick and ban rate per hero and per rank; win-rate shift for each pair of heroes, as enemies and as allies.
  • Official, but a rule rather than a measurement: team needs. Moonton publishes each hero’s classes, specialities, Offense and Durability ratings and the damage type of every skill. From them we read whether your team — locked allies, plus what the open lanes usually bring — leans on one damage type or lacks a frontline, and add a few points to a hero that fills the gap. The hero profiles are Moonton’s; the idea that a one-sided team is weaker is a rule of this tool — the pair figures do not show it — so the bonus stays small, never subtracts, and fades when the enemy draft counters the hero.
  • Measured figures, read under a convention: counter-pick risk. For each hero and each enemy lane we average its measured matchups against the least favourable quarter of that lane’s picks: what the hero stands to lose if the enemy picks after it. That figure is stable from one rank to the next. It is compared with the other heroes of the same lane, so it reorders picks without lifting every score, and it is only counted for enemy lanes still open: once the enemy has locked a hero, the real matchup is read instead. Assuming the enemy will pick a counter is a convention of this tool, so the weight is kept low (under one point today). A hero picked too rarely for its matchup figures to be reliable gets no adjustment at all. On an empty draft, about six of the seven recommended heroes of a lane stay the same as without it, and the first pick changes in about one lane in three.
  • Not scored: enemy composition profiles and hand-written signature combos. We tested hero types (class, speciality, durability, damage type) against the pair figures: they explain about 5% of what the measured matchups already say, so a rule built on them would count the same thing twice. Signature bonuses written by hand would be opinions, not data.
  • We read pair figures as team figures, not lane duels. Moonton does not document their scope; our own measurements point to a shift that holds whichever lane each hero plays. All five enemies are therefore read at the same weight.
  • Sign convention. Moonton does not document it. We read a positive shift as favouring the first hero of the pair, which matches Moonton’s own counter lists in about six cases out of ten — a weak check, set out on the method page.

How the score is built

Moonton publishes, for each pair of heroes, how far a hero’s win rate moves from its own average when the other hero is in the match. Each figure is a shift from the hero’s average, so the shifts of the locked heroes add up; that sum is what moves the score.

  • Every locked hero counts, enemies and allies alike, and a negative shift counts as much as a positive one.
  • Rare pairs count for less. Moonton gives no match count per pair, so we measure the noise of each figure from how much it moves from one day to the next: it grows as the two heroes are picked less. A pair is counted in proportion to the share of signal in its figure, instead of being trusted in full or thrown out.
  • Checked across ranks. Ranks are separate groups of players. A rank’s weighted pair figures predict the next rank’s published figures better than its raw ones in every case we tested, most of all for rare pairs.
  • The hero’s own share is a setting, not a measurement. Ranked on raw win rate, one rarely picked hero sat in the top seven of almost every draft of its lane (98.5% of simulated EXP-lane drafts at Mythic, for 0.37% of picks, measured on 14 November 2026), and up to half of a lane’s heroes showed up in fewer than 2% of drafts. The hero’s share is therefore tempered by how often it is picked, with the same rule as our tier list, and enemy pairs count 1.3 times. Both settings were chosen so that the locked heroes, not one statistic, decide the ranking.
  • Checked on thousands of simulated drafts. At each of the six ranks we verify that no hero sits in more than 90% of the top-seven recommendations of its lane, that heroes winning at least as much as their lane do not vanish, and that a clearly bad enemy pair pushes a hero out of the top picks. These checks say the tool is balanced and consistent; they do not say its picks win.
  • The build shown under a pick is not scored either. It is the hero’s most-played build in that lane at the chosen rank, as Moonton records it: three core items, an emblem with its talents and a battle spell — not a full six-item build, not a purchase order, and not specific to the draft on the board. Its win rate is left out on purpose: Moonton lists only a hero’s three most-played builds per lane, and those win about five points more than the hero does overall (measured on 15 November 2026). The share shown is the part of the hero’s games in that lane played with this build.
  • The composition panel is a reading, not a score. Under each team it shows the average of Moonton’s official Durability, Offense and Ability Effects ratings (0–100) for the heroes picked, with a tick for what a usual team at that rank reaches (each lane’s pick-weighted average), the physical/magic split of their damage weighted by Offense, and the frontline count. “Lacks” applies the team-needs rule to the whole team, open lanes counted for what they usually bring. The three averages shown are not counted in the score; the team-needs rule itself reads Offense and Durability hero by hero, as described above, so a recommended pick can carry a needs bonus for its own lane that this team-wide line does not flag.
  • Team needs are checked the same way. With the four other lanes locked and no frontline among them, at least two frontline heroes reach the top five in 99% of simulated Mythic drafts (74% without the rule, measured on 14 November 2026). A lane still open counts for what it usually brings: Moonton lists only mages in the mid lane, so a team is not called short on magic damage while its mid is unpicked, and in practice the damage-mix bonus mostly helps physical heroes join a magic-heavy team. A team that is already balanced gets exactly the ranking it had without the rule.

Two assumptions are built in and cannot be tested without match records: that day-to-day movements are independent, and that pair effects add up. The score says what Moonton’s statistics imply under those settings; no record of real drafts and their results exists to test it against.

How to use it well

Compare the first three picks and read why each one is there, instead of locking the first.

  • Give more weight to a pick whose three bars are read than to one ranked on hero stats and counter-pick risk alone.
  • In ban mode, once heroes are locked, the row turns into suggested bans: Moonton’s ban rate, plus how much each hero would gain against your locked heroes and alongside the enemy’s, from the same measured pair figures. Blending the two is a rule of this tool. A hero the enemy has no lane left for is never suggested.
  • A rarely picked hero has noisy pair figures at that rank: they count for little, it is ranked mostly on its own win rate, and its breakdown says so.
  • The numbers know nothing about your own mastery of a hero. Cross-check with the tier list, the counter desk and each hero page.

Ranks, bans and data windows

The rank selector switches the whole data set: Epic, Legend, Mythic, Mythical Honor, Mythical Glory and above, or all ranks together. Lower ranks have no pair figures and are not offered.

  • Hero stats cover 7 days, pair stats 15 days. Both windows and dates are shown under the board, with a warning when they are more than a week old.
  • Bans only remove heroes from the recommendations. Up to five per team can be entered; the tool does not enforce the number your rank uses.
  • The tier letter on a portrait is the hero’s standing within its lane at the chosen rank, by the same method as the tier list.

Draft tool FAQ

Is this an official Moonton tool?

No. MLBB Core is an independent fan project. The tool reads statistics that Moonton publishes and is not affiliated with or endorsed by Moonton.

Why does a recommended hero have a bar marked “not read”?

Because no figure exists for that channel in the current draft: no enemy or ally is locked yet, or Moonton publishes no pair figure for that hero at this rank. The bar then shows “not read”. One case still moves the score, and the bar says so: with an ally locked, a hero with no published pair is scored slightly below its average partner.

Does the top pick guarantee a win?

No. The score ranks heroes by published win rates in similar team contexts. It does not know your skill on a hero, and it has not been tested against real draft results, because no such record is available.

Can I share a draft?

Yes. The Share button copies a link that restores the lane, the rank, both teams and the bans.