The International 2024 (TI13) ended with Team Liquid lifting the Aegis after seven long years. While we became only the third organization after OG and Team Spirit to win the holy grail of Dota 2 twice, it was the first time ever a team did it with a completely different roster. 

For Liquid, the win feels more hard-earned than ever. Call it a silver medal curse or a Grand Finals demon, this core has come extremely close to winning multiple trophies. Team Liquid exorcized their demons and curses in Peru earlier this year when they won Elite League Season 2, and maybe it was that confidence that let them clean sweep their biggest rivals, Gaimin Gladiators, in the TI13 Grand Final.

No one wins on confidence alone, though. In Pro Dota, data seems more important than ever in the last few seasons, with teams employing more coaches and analysts to pour over the numbers. So, what do the numbers say about the nitty gritty of TL’s win? For many Team Liquid fans, the main stat is the unprecedented 100% win rate in the TI13 playoffs. 

But if you want to understand the “why” behind that 100% win rate, you have to get more intricate. For that, Team Liquid has SAP. Based on the data provided by SAP, we were able to take an in depth look at the numbers behind TI13 and pinpoint to the ones that set Team Liquid apart.

The sweet spot: Unique Heroes Played

Our first data stop is at unique heroes played at The International 2024. The bar graph below shows the unique heroes which  each of the top 8 teams used in the pursuit for the top spot, with the number of games played by each team provided in brackets besides the team name. Team Liquid fall somewhere in the middle for the number of games played in the tournament, yet their unique hero count is nothing to brag about. This says a lot about the optimum number of strong heroes a team requires in order to beat the competition.

The Captains Mode format for the current Dota 2 patch has 7 bans and 5 picks for each team. A total of 24 heroes play a part in the draft for a game. The first phase bans are mostly the same bunch of heroes that are extremely strong in the meta, while the second and third phase bans concentrates on the team and the draft specifically. So even with 19 heroes out of the game through bans and opponent picks, it still leaves a team 15 strong heroes to pick from (with the number 34 in mind).

Liquid’s win is a testament to the fact that there is a sweet spot in the number of heroes you display in your arsenal. Unless you are the Wings Gaming roster that won TI6, attempting to pick too many heroes can be detrimental.

Gaimin Gladiators, who had the highest number of unique hero picks, were one series win away from proving that narrative incorrect. But even Tundra Esports, who finished third at TI3, did not have a very high number of unique heroes to their name.

A good example of the impact of this is the TI13 grand final. Team Liquid first picked Nature’s Prophet in all three games after completely ignoring the hero in the two prior series. Gaimin Gladiators did not have it on their ban priority list, and tried a myriad of strategies to defeat the first pick Nature’s Prophet, but to no avail. Sometimes it just takes one good pick. 

‍

The miCKe funnel: Farm Distribution

The farm distribution for the top 8 teams on the basis of GPM and last hits is shown below for all games before the last 3 days, which were played in the Royal Arena. It is a very interesting distribution for Team Liquid.

Cores

Amongst the cores, Michael “miCKe” Wu had the highest farm priority, which is not surprising. All of the top 8 carries had the highest farm priority in their team. The adjustment that can be seen is from Michał “Nisha” Jankovski and Neta “33” Shapira. When 33 won The International with Tundra Esports, he was basically playing position 2 from the offlane, and was one of the highest farm priority offlaners. But coming to Team Liquid, he modified his playstyle to fit his new team. The same could be said about Nisha, who was playing the position 1 role from the mid lane back in his Team Secret days, but has now become adept at playing tempo controlling heroes like Puck and Tiny until miCKe comes online.

Supports

In both parameters used for looking at the farm distribution, Samuel “Boxi” Svhan had a lower farm priority than Aydin “Insania” Sarkohi, although on paper, Boxi is the position 4 player. Team Liquid was the only team with this kind of farm split, and it shows that both supports are comfortable playing position 4 role or position 5. This used to be quite common, but in recent years, the roles have become more rigid. Liquid’s fluidity showed in more than farm, but also in that they switched lanes in the laning stage when required.

Top of the stack: Neutral Camps Stacked

Neutral camp stacks are a fascinating indication of the non-linearity of Dota 2. Stacks early on in the game for the core heroes can help them get to a key item like Black King Bar (BKB) or Blink Dagger a little early, which can completely change the direction of the game. In terms of the net worth difference, stacks might have a small impact, but they remain to be a key tool for cores to hit their item timings.

Amongst the top 8 teams, Team Liquid had the second highest average stacks at around 18.3 stacks per game. 33 was the highest stacker amongst offlaners and Insania was second amongst position 5 players. But the thing that set Team Liquid apart from most of the other teams is that all of their players were involved in making stacks, and it was not something left just to the supports. 

The gold advantage for a team having 10 additional stacks can’t be more than 1720 gold, and that too would only be the case if all 10 of those stacks are ancient stacks. The impact is based on when that extra gold is made available, and Team Liquid definitely made the most of it.

Objective-based approach: Building Damage

Dota 2 is a complex game, but when push comes to shove, it’s about destroying the enemy buildings. You take enemy towers, barracks and Ancient, and then you win! This is all to say building damage constitutes an important parameter in the game, and two of Liquid’s players stood out. Nisha and 33 had significantly more damage than their positional counterparts. Nisha had 61% more building damage throughout the tournament compared to the mean of top 8 mid players, while 33 had a whopping 142.9% more building damage than the mean of the top 8 offlaners.

A point can be made that Team Liquid only dropped 3 games out of the 18 games they played at TI13, and naturally, winning teams will have higher building damage. While that may be the case, the percentage difference seems too high for that to be the only factor. Other factors include hero picks like Nisha on Nature’s Prophet or Tiny and 33 on Beastmaster and Visage, as well as how the team divided the map and delegated lane pushing and tower hitting responsibilities.

Playing together: Assist Efficiency

All five Team Liquid players had very high kill involvements (Kills+Assists) at TI13. But that data can get skewed for teams that do well in tournaments and have a very high win rate. The question is, what did the assist efficiency look like?

Assist efficiency is a parameter that indicates how involved the team is in kills. A team can have 20 kills, and the assists can range from 0 to 80 – one extreme being a single hero got all the kills and the other extreme being all five heroes were involved in all kills. In that case, 80 assists would be an assist efficiency of 100%, indicating that the team had great cohesion playing together.

For the top 8 teams at TI13, Team Liquid had the second highest assist efficiency after Xtreme Gaming, just shy of 70%. The numbers here are not significantly different, with the highest and lowest teams separated by less than 9%. But when the best in the world fight it out, the margins are extremely narrow, and every small detail makes a difference.

No snowballing allowed: Net Worth Difference at 10 Minutes

The laning stage dictates a lot of how a game of Dota 2 is played out. We looked at the Laning Stage Behemoths for TI12 at the end of The International 2023, which showed how dominant Gaimin Gladiators (and to a lesser extent, Team Spirit) were in the laning stage. Gaimin Gladiators have made a name for themselves for being laning gods, and even in a meta that favors long games, they will find a way to roll opponents over in the early game and take the base a lot quicker than most teams. That is their modus operandi, and they did it to an extent at TI13 as well.

The table below shows the average net worth difference at 10 minutes for the top 8 teams. Gaimin Gladiators dominated this stat last year with an average net worth advantage of 1544 at TI12, but came in second to Team Falcons this year.

Team Liquid did not record a crazy net worth advantage at 10 minutes for their TI13 games. But their playstyle does not rely on being ahead so early in the game. All they had to do was ensure that they did not fall too far behind. In all 5 games Team Liquid played against Gaimin Gladiators at TI13, they were trailing in net worth at the 10 minute mark. The average net worth deficit was 1003 gold and Liquid’s biggest deficit at 10 minutes was only 1370 gold. But that is all Liquid needed to stay in the game until their heroes came online. 

In comparison, Tundra Esports had an average net worth deficit of 2446 gold at 10 minutes in their two upper bracket semifinal games against Gaimin Gladiators. That is approximately the cost of five Bracers, which early on in the game can make all the difference in team fights.

‍

On the same side: First Pick Win Rate

One of the most skewed metrics at The International 2024 was the first pick win rate. In the playoff games, first pick had a win rate of 58.8%, which increased to 63.6% when the data for the final 3 days played in the Royal Arena was analyzed. This is due to a combination of factors – the ban and pick phases in Captains Mode for Dota 2 patch 7.37, the strong heroes in the current Dota 2 meta as well as the lack of benefit that either Radiant or Dire provide.

How does this factor into Team Liquid’s TI13 win? The Boys in Blue did themselves a huge favor by ensuring they reached the grand final through the upper bracket. In all series, the choice of side and first pick is decided through a coin toss. For the grand final, the team that comes through the upper bracket gets to make a decision in Game 1. Team Liquid chose first pick, won the game, got the momentum on their side and ended up winning 3-0. They also got to choose first pick in Game 3. In Game 2, they defied the odds to win with second pick.

The data provided by SAP shows that Team Liquid was not the absolute best team in most of the measurable categories, but they were above average in almost all metrics. And that is what the players, coaches and staff of great teams do – ensure the team is well-balanced and considers every small detail that could make a difference. 

While SAP may have provided us with some deep insight into the TI13 statistics that gave Team Liquid the edge, there is a final SAPpy metric that cannot be measured by data collection from the Dota 2 games. That's camaraderie. The core of this Dota squad has stuck together for years through thick and thin, gradually climbing their way to the top. Amidst retirements and curses and rivalries, they remained loyal to one another and became so in-sync and fundamentally solid that they eventually stormed their way to an Aegis. 

‍