Dota 2 Match Analysis: Drafts, Patches and Best-of Formats

Dota 2 is the esport where the most important decisions are made before the first creep spawns. The draft — ten picks and a long list of bans, taken in phases by two captains — shapes a game more than any single play in it, and a balance patch can rewrite what a good draft looks like overnight. Add a tournament calendar that switches between best-of-two group stages, best-of-three playoffs and best-of-five finals, and analysing a Dota 2 series becomes an exercise in knowing which rules apply today. This guide covers the formats, the draft, the patch cycle, rosters, the market line where a public book prices the series, and how Sportslyx grades the result.

Three formats, three different questions

Dota 2 uses more series formats than any other esport on the Sportslyx board. Group stages at majors and qualifiers are often best-of-two, where a 1–1 result is a legitimate outcome and the standings are decided on points. Playoffs are best-of-three. Grand finals are best-of-five, and tiebreakers are sometimes a single game. The format is listed with every series on the Dota 2 matches page, and it should be the first thing you read.

The best-of-two deserves special attention because it has no equivalent in most sports. Two evenly matched teams split it more often than not, and a favourite that wins the first game has already banked a point and may draft experimentally in the second. For analysis, treat a best-of-two as two games rather than a series with a winner: the questions are whether the favourite sweeps and whether the underdog can steal one. For grading, a 1–1 best-of-two has no series winner, and Sportslyx does not score it as a win or a loss — a read on it is voided rather than guessed. The one-map push described later exists only in best-of-three and best-of-five series.

The draft you cannot see yet

Everything you analyse before a series is done without the one thing that decides it: the draft. Captains ban and pick in alternating phases, first pick and last pick carry different advantages, and the final composition determines whether a team is playing for an early lead or a late one. You cannot know the draft in advance, but you can know how a team drafts.

• Comfort picks: heroes a team runs whenever they are available, and which opponents ban as a matter of routine
• Flexibility: a roster whose cores can play several roles is harder to ban out than one built around a single signature hero
• Draft tendencies: some captains draft for lane dominance and short games, others for scaling and long ones; the tendency predicts how a game is likely to go even when the specific heroes do not
• Side selection: the team that wins the coin toss chooses between first pick and side, and the choice tells you which it values on the current patch

The practical rule is to read a team’s last several drafts, not its last several results. A team that has been winning with one strategy on one patch is exposed the moment the patch changes or a captain prepares a counter; a team that has shown three different ways to win is a sturdier favourite at the same price.

A new patch resets the form table

Dota 2 is balanced by patch, and a large patch is the closest thing in esports to a new season. Hero strength shifts, item timings move, and strategies that dominated for months can disappear within a week. Results from before a major patch are evidence about a different game. The first tournaments after a patch are the noisiest on the calendar — teams are still discovering what works, upsets are common, and a favourite’s price should carry more doubt than the same price a month later.

Tournaments usually freeze the patch in advance, so check which version a series is played on and how long the teams have had with it. A team that adapts quickly is worth more in the first two weeks of a patch than its record shows; a team whose whole identity was built on a weakened strategy is worth less. The Sportslyx breakdown reads recent form and the tournament context, but it does not know which patch a team practised on — that is the analyst’s job.

Rosters, shuffles and the reserve feed

Dota 2 rosters change more than those in any other esport, and the changes cluster: the weeks after The International are a shuffle season in which a large share of top teams replace at least one player. A new roster’s early results are small samples with a fresh set of tendencies, and head-to-head history from before the change belongs to a different team. Sportslyx reads rosters and crests from bo3.gg’s public feed along with the schedule, live scores, results and the tournament and best-of format; for Dota 2 there is also OpenDota as a reserve, which supplies live games and recent results when the primary feed fails — though not the upcoming schedule, so a day served from the reserve says so on the board.

• Stand-ins: a team playing with a substitute is not the team that produced its recent record, and the draft usually narrows to what the stand-in can play
• Regional strength: results inside a weak region inflate a record against the world
• Recent map scores: 2–0 wins against peers say more than 2–1 wins against qualifiers

Reading the series line

Where a public bookmaker covers the series, Sportslyx reads three markets from a free public feed: the series moneyline, a map handicap and a total-maps line. In a best-of-three the map handicap is usually ±1.5 — the favourite at −1.5 needs a 2–0, the underdog at +1.5 covers by taking a game — and the total is 2.5, which is a question about whether a third game is played. In a best-of-two the total-maps line, where it exists, is effectively a question about a 1–1 split.

The moneyline becomes the win chance on a Sportslyx page through a fixed calculation. Take a series priced −135 and +110. The implied chances are 57.4% and 47.6%; they sum to 105.1%, and the 5.1 points above 100 is the bookmaker’s margin. Dividing each figure by the sum gives 54.7% and 45.3% — a near coin flip once the margin is gone, which is how many Dota 2 playoff series look. That market-implied figure is what the public page shows; it is a statement of what the market believes, not a Sportslyx pick.

Game length and the total-maps market

Dota 2 games are long — commonly thirty to fifty minutes, sometimes more than an hour — and a best-of-three that goes the distance can run past two and a half hours. That matters for two reasons. First, the total-maps line is a read on closeness, not on quality: a strong favourite can go 2–1 against an underdog with a good first draft, and an even series can end 2–0 on two lopsided drafts. Second, a team playing its second long series of the day is a different proposition from one that swept its morning match; fatigue is real in a game this long, and the board’s start time is worth checking against the previous series.

How Sportslyx grades a Dota 2 read

Board reads are graded from the final map score. The chosen team winning the series is a win. A series lost by exactly one map — 1–2 in a best-of-three, 2–3 in a best-of-five — is a push, returned rather than counted as a loss, and the rule is stated on every esports page because it lifts the published hit rate relative to a plain win-or-lose count: no public book settles a series moneyline with that insurance. A 0–2 or 1–3 is a loss. A best-of-one has no push, and a tied best-of-two is voided.

Reads are selected by expected value with a price floor: the favourite’s calibrated series chance — calibrated because games within a series are correlated and a raw series probability overstates a favourite — is set against the available price with the push treated as a returned stake, and a series only qualifies if the expected return clears a fixed threshold. A favourite priced too short is passed over, and formats without a push rarely clear the bar. Hit rate is wins divided by wins plus losses, pushes excluded; ROI is at a flat stake and labelled “partly est. prices” because some priced series carried a model estimate rather than a book line. The record is published by calendar month in UTC on the Dota 2 predictions page and the Pick Tracker.

A checklist before the series

1. Read the format first — best-of-two, three or five — and decide which questions apply
2. Check the patch and how long each team has played on it; discount results from before a major patch
3. Confirm both rosters match the ones behind the recent results, especially in the weeks after a roster shuffle
4. Read the market-implied chance where a book line exists, and treat an unpriced qualifier as unpriced
5. Look at the last several drafts for comfort picks, flexibility and side preference rather than at the last several scores
6. Log your own read before the scheduled start — tracked picks close fifteen minutes after it — and judge the method on a month of graded series

Summary

Dota 2 analysis is the discipline of knowing what has changed: the format, the patch, the roster and the draft tendencies that survive all three. On Sportslyx the win chance is market-implied where a public book prices the series, the AI breakdown is written around that number and the recent form, and every board read is graded from the final map score with a one-map loss counted as a push. The Dota 2 analytics hub has the current board, the methodology fixes the grading rules, and the esports board runs the same method on every series of the day. Past results do not predict future results, and no outcome is certain.