PARI Universe Playoffs Preview: Teams, Bracket, Draft Data

Final Group Standings

Ten teams, two groups of five, a single round robin, every series Bo3. Each team played four series. We label the groups A and B by the teams that played each other, so check the official bracket for FISSURE’s own naming.

Group A

Group B

The top of both groups is clean. Team Yandex and Team Spirit both finished 4-0 with an 8-1 game record, and each dropped exactly one game, Yandex to Team Liquid and Spirit to PARIVISION. PARIVISION and GamerLegion took the second spots at 3-1. Everything interesting happened below them.

Team Liquid are the biggest story in Group A. After a 0-2 loss to GamerLegion they pushed both OG and Yandex to a deciding third game, so a 1-3 series record undersells how close they were. We covered how that roster came together in our Malr1ne joins Team Liquid breakdown.

The Group B Three-Way Tie

Group B finished with a knot. Astra Invicta, BetBoom Team and LEGION all ended 1-3, and the head-to-head results form a perfect circle: Astra beat BetBoom 2-0, LEGION beat Astra 2-1, and BetBoom beat LEGION 2-0. Nobody can separate them by results against each other.

If the tiebreaker falls to game differential, which is the normal next step, the order is clear:

• Astra Invicta: 3 games won, 6 lost, a difference of -3
• BetBoom Team: 2 games won, 6 lost, a difference of -4
• LEGION: 2 games won, 7 lost, a difference of -5

On that basis Astra Invicta and BetBoom Team take the last two Group B playoff places and LEGION go home. We want to be straight about the limit of this claim: we calculated the game differential ourselves and could not find FISSURE’s published tiebreaker rules in anything we could verify. If the organizers use a different method, the answer could change, so treat LEGION’s elimination as our calculation and not an official ruling.

It is also worth understanding why this tie happened. LEGION’s only series win came against Astra on October 24: they took game one in 36 minutes, lost a 54-minute game two, and won game three in 48 minutes. If Astra had closed that series in two games, the circle would never have formed. One game decided who is in the bracket and who is on a plane home.

Playoff Format, Schedule and Prize Money

The playoffs run October 28 to November 1 as an eight-team double-elimination bracket. Every series is Bo3 except the Grand Final, which is Bo5. That format punishes a slow start less than the group stage did: one lost series drops a team to the lower bracket instead of ending its event.

The posted schedule slots, in UTC, look like this. Dubai is UTC+4, so add four hours.

The prize pool is $250,000 total. The listed splits are $125,000 for first, $50,000 for second, $30,000 for third, $14,000 for fourth, $10,000 each for fifth and sixth, and $5,000 each for seventh and eighth. First place alone is half the pool, so every team in the field is playing for a result with real separation between the top two and everyone else.

Team-by-Team Playoff Read

All numbers here are group-stage games on patch 7.41f. We are reading tendencies, not predicting winners. Pro Dota swings hard from series to series, and 4-9 games per team is a small sample.

Team Yandex (4-0, 8-1 in games)

The cleanest group record of any team, and they already hold a reputation: they beat Team Spirit in the BLAST Slam VIII final, which we covered in Yandex beat Spirit 3-1. In this group stage their signature picks were Earth Spirit, Techies and Windranger, three games each. Their average game length was 40.6 minutes, close to the field’s 42.8 average, so they are not a team that needs the game to be short or long.

Their most dramatic game was the 71-minute win over OG on October 24 (match 9059530192), a 71-34 kill score in Yandex’s favor where watson carried Ursa to 22 kills and 21 assists at 885 GPM. That is a team that can win from a long, messy late game and also finish a 24-minute stomp against 1w Team.

Team Spirit (4-0, 8-1 in games)

Spirit were the other perfect group record, and their drafts were the most distinctive. Hoodwink five times, Witch Doctor four times, Sand King four times, Slark three times. Larl’s Sand King (4 of 9 games) and not_me’s Hoodwink (4 of 9) were near-signature, and Yatoro on Slark produced a 17-0-12 line against LEGION in match 9056429802.

Their one lost game came against PARIVISION on October 26: a 48-minute game where Satanic’s Nature’s Prophet finished 20-2-8 at 994 GPM. Spirit won game one in 32 minutes and won the deciding game three in 32 minutes as well, so the loss in between was the only blemish on their group record.

PARIVISION (3-1, 7-2 in games)

PARIVISION only lost to Spirit, and they took that series to a third game. Their identity is built on Dukalis on Witch Doctor (5 of 9 games), Noticed on Brewmaster (3 games) and Satanic on Kez (3 games). They had the shortest average game of any team at 37.0 minutes, so they tend to convert leads quickly instead of dragging games into the 60-minute zone.

Satanic is the player to watch. He posted three of the top six GPM lines in the whole group stage, on Nature’s Prophet (994), Templar Assassin (983) and Terrorblade (946). A carry who can win on three very different heroes is hard to ban around.

GamerLegion (3-1, 6-3 in games)

GamerLegion beat Team Liquid, 1w Team and OG and only lost to Yandex. They are the team most tied to a single hero: Bignum played Dark Willow in 6 of his 9 games, and Dark Willow went 11-7 across the whole tournament. RCY added Dragon Knight four times and Invoker three times. Their average game ran 46.9 minutes, so they live in long games.

Their comeback in the 70-minute game against Team Liquid on October 23 (match 9057872455) came after trailing by about 7,900 gold at the worst point. That is the kind of patience a double-elimination bracket rewards.

OG (2-2, 5-5 in games)

OG finished with an exactly even game record, but they own the single largest comeback of the tournament. Their average game length was 51.0 minutes, the longest of any team. In match 9060951574 they beat Team Liquid after trailing by roughly 15,500 gold at their lowest point.

The draft story: skem played Techies and Dazzle three times each and SSS played Timbersaw three times. Those picks lean on the late game, which fits a team whose games run long. Our read is that this is a risk against a team like PARIVISION, whose games average just 37 minutes.

Team Liquid (1-3, 4-6 in games)

The fourth-place team in Group A was closer than its record. They pushed Yandex and OG to three games each, won the longest game of the tournament (76 minutes against 1w Team), and played Malr1ne on Earthshaker in 4 of his 10 games, tOfu on Witch Doctor four times, and Boxi on Spirit Breaker three times. Their average game was 47.9 minutes, second longest in the field.

Liquid’s issue in this stage was converting. They lost the deciding third game both against OG on October 25 and against Yandex on October 26. In a Bo3 playoff format that is the difference between the upper and lower bracket, so draft discipline in game three will matter for them.

Astra Invicta (1-3, 3-6 in games)

Astra were the biggest question mark in Group B. Their plan is clear from the data: Nightfall on Lifestealer in 4 of 9 games and lorenof on Ember Spirit in 4 of 9, with GH on Io three times. They beat BetBoom Team, but lost to PARIVISION, Spirit and LEGION. Lifestealer and Ember Spirit are both heroes that want space to farm, so our read is that Astra need a slow opening to be at their best.

BetBoom Team (1-3, 2-6 in games)

BetBoom got in on game differential. They played an unusual mix: Save- on Keeper of the Light (3 games), Kataomi on Witch Doctor and Hoodwink (3 games each) and MieRo on Brewmaster and Doom (2 each). Their 2-6 game record was bad, but they averaged just 37.8 minutes per game, which suggests their games were decided fast and mostly against them.

Eliminated: 1w Team and LEGION

1w Team went 0-4 in series and 0-8 in games. Their closest moment was the 76-minute opener against Team Liquid, the longest game of the tournament, which they lost 47-41 in kills. LEGION finished 1-3 in series and 2-7 in games, and on our calculation are the odd team out of the Group B tie.

Draft Data From All 45 Games

Eighty-five different heroes were picked and 74 were banned across 45 games, which tells you this is not a one-hero meta. But a handful of heroes appear again and again. Here are the twelve most-picked heroes in the group stage with their results and how often each got banned.

A few things stand out right away.

• Witch Doctor was the most-picked hero in the tournament, 22 times, with a 14-8 record (64%). Six of the ten teams played it, led by PARIVISION (5 games), Spirit (4) and Team Liquid (4).
• Dark Willow went 11-7 (61%), and two players leaned on it hardest: GamerLegion’s Bignum (6 games) and LEGION’s Danial (4 games).
• Hoodwink and Dragon Knight were popular and weak. Hoodwink went 7-12 across 19 picks and Dragon Knight went 6-10 across 16 picks. High pick rate with a sub-40% win rate usually means the hero fits many drafts but is easy to play around.
• Dazzle went 3-9, the worst result of any hero with double-digit picks.
• Techies was picked 20 times and went 9-11, a near-even record in games that averaged 44.5 minutes, a little longer than the 42.8-minute tournament average.

The Ban Phase Tells the Story

Every hero on this list was banned at least 19 times in 45 games. These are the heroes teams were most afraid of, or most wanted to deny each other:

Bounty Hunter was banned in 40 of 45 games (89%) and was still picked in 5 games, winning 3 of them. That is the kind of hero teams never leave open. Enigma was banned 36 times and picked only 5 times. Slark was banned 32 times and picked only 6, going 4-2 when allowed.

The more interesting entries are Lone Druid, Brewmaster and Pangolier. Lone Druid (30 bans) went 6-3 when picked, Brewmaster (29 bans) went 6-3 and Pangolier (18 bans) went 6-2. These are heroes that got heavy ban attention and still won at least two-thirds of their games when they slipped through the ban phase.

The practical takeaway for your own games: a pro ban list reflects what a coordinated team fears from a specific opponent, not what is dangerous at every rank. Do not copy pro bans in ranked, and do not assume a hero is weak just because it is rarely picked at your level.

Marathons, Comebacks and Stomps

The average game lasted 42.8 minutes and the median was 37.9, so there is a long tail of very long games pulling the average up. Seven games went past 60 minutes and five finished in under 30. Radiant won 22 of 45 games, which is as close to a coin flip as you can get, so side selection is not a factor in the playoffs.

The five longest games of the group stage:

The longest game, Team Liquid over 1w Team at 76 minutes 59 seconds, is also the most remarkable. Liquid trailed by around 11,400 gold at the low point, ended up winning 47-41 in kills, and later beat the same opponent again 36-7 in 26 minutes. Same two teams, and a 50-minute gap in game length between the two games. Draft and early momentum shape game length as much as team strength does.

The 71-minute Yandex win over OG produced the most kills of any game with 105 (71-34). Three other games went past 65 minutes: the 76-minute Liquid game, the 70-minute GamerLegion win over Team Liquid and the 67-minute OG win over Team Liquid. Team Liquid played in three of the four longest games of the group stage.

Standout Individual Performances

Some individual lines worth looking up in the replay:

• watson (Team Yandex), Lone Druid: 16 kills, 1 death, 15 assists and 1,043 GPM, the highest GPM of the entire group stage, in a 61-minute win over OG (match 9059391986).
• Satanic (PARIVISION), Nature’s Prophet: 20 kills, 2 deaths, 8 assists, 994 GPM in the game that took Spirit to a third map (match 9063230342).
• Yuma (GamerLegion), Lone Druid: 18 kills, 1 death, 9 assists, 985 GPM against OG (match 9062644050).
• watson (Team Yandex), Ursa: 22 kills, 7 deaths, 21 assists in the 71-minute game against OG (match 9059530192), the most kills by any player in a single game.
• m1CKe (Team Liquid), Windranger: 20 kills, 3 deaths and 14 assists in the 76-minute win over 1w Team (match 9056618744).
• Yatoro (Team Spirit), Slark: a 17-0-12 line, no deaths, in a 33-minute win against LEGION (match 9056429802).

The pattern: carries who farm fast on a hero banned in most drafts (Lone Druid, Slark, Nature’s Prophet) turned in the biggest numbers. Those heroes appear on ban lists for exactly that reason.

What to Watch in the Playoffs

Here is our read, labeled as analysis and not as fact.

1. Game length mismatches. PARIVISION (37.0 minute average) against OG (51.0) or Team Liquid (47.9) is a clash of tempos. Whoever controls the draft controls the length.
2. Witch Doctor and Dark Willow contests. Both were top-four picks with winning records. Expect them to either get banned early or be first-phase picks in the playoffs.
3. Whether Bounty Hunter gets let through. At 89% bans, a playoff game where Bounty Hunter is open means a team made a deliberate trade somewhere else.
4. Team Liquid’s third-game record. They lost 1-2 to OG and 1-2 to Yandex. Bo3 playoff series need either an early lead or a calm draft.
5. The lower seeds from Group B. Astra and BetBoom reach the bracket after losing three series each. Their opening series is the most likely place for an upset, because both teams are playing with nothing to protect.

For the wider context on who the best teams in the 2027 season are, see our post-shuffle power rankings, and for the next big event after this one, our DreamLeague Season 30 preview.

What Ranked Players Can Steal

Pro drafts do not translate directly to ranked, but a few lessons do.

1. Position 5 Witch Doctor is still a strong, simple pick

Witch Doctor was picked 22 times and won 64% of them. The kit is easy to understand and rewards good positioning more than mechanics. If you play position 5, it is one of the lowest-risk heroes to put time into.

2. Match your draft to your game length

Games here averaged almost 43 minutes. In our experience coaching and boosting, ranked games below Divine often finish earlier, so early tempo matters more than the pro data implies. Do not copy a draft built for a 60-minute game if your games end at 30.

3. Heroes with huge ban rates win when allowed

Lone Druid, Brewmaster and Pangolier all won at least two-thirds of their games despite 18 to 30 bans each. If your opponents leave one of them open, play it with a plan, because the tournament results suggest it punishes teams that do not have one.

4. Popular is not the same as strong

Hoodwink (7-12) and Dragon Knight (6-10) were two of the six most-picked heroes and finished well under 50%. If you copy a pick because you saw it in a pro game, check the results first. Our hero win rates by bracket and hero tier list are better references than a single tournament.

5. Coaching and boosting are different tools

Watching pros helps you spot patterns. Applying them to your own replays is faster with a coach who sees what you actually do. If you would rather get the rank first and learn from the games afterward, our guide to the best Dota 2 boosting services explains how to compare providers, our safety guide covers account risk, and our MMR boosting guide explains how the process works. You can also go straight to Dota 2 coaching or the MMR boost page.

Full Results Table: All 45 Games

Every group-stage game in order of start time. Click a match ID to open it on OpenDota.

Data Notes

• All hero, player, kill, duration and gold numbers come from OpenDota match data for league ID 20395, 45 games, patch 7.41f, collected on the morning of October 27, 2026.
• Group-stage standings, format, schedule and prize splits are from the public tournament page on hawk.live. We did not verify the official seeding of the playoff bracket.
• The Group B tiebreak is our own calculation of game differential. FISSURE’s official tiebreaker rules were not available to verify.
• Hero win rates here are tournament-only and based on small samples. A hero with 5 picks can look dominant or terrible with one game.
• Player names are as shown in the OpenDota data. Where a name contains non-Latin characters we use the Latin part only.
• “Gold deficit” is the largest gold disadvantage recorded for the eventual winner at any minute of the game.

FAQ