We Analyzed 584 Dota 2 Games: The 3 Mistakes That Cost Players the Most

Every DotaMirror report ends with one line: the single most important thing to fix before your next game. After six months and 584 saved reports, we went back through all of them to see what that line says most often.

The short version: almost every game comes down to one of three things. You died too much, you built the wrong items for the game in front of you, or you farmed too slowly. Here is what the data shows for each, and what to actually do about it.

What We Looked At

The data set is every coaching report saved by signed-in DotaMirror users between 30 April and 29 October 2026: 584 games across 106 different heroes. 87% were Ranked All Pick and 12% were Turbo. 53% of the games were wins and 47% were losses, so this is not just a collection of people's worst games.

A few honest caveats before the numbers. Only reports from signed-in players are stored, so anonymous analyses are not included. The players who use a coaching tool are not a random sample of the whole Dota population. And the themes below were grouped automatically from the report text, so treat the percentages as a good approximation rather than an exact count. Nothing in this post identifies a player or a match.

The #1 Priority, Grouped by Theme

Three themes cover 82% of all reports. Laning, warding and teamfight presence, the things most guides spend the most time on, almost never turn out to be the single biggest problem. They matter, but in most games something more basic is costing more.

Mistake 1: Dying Too Often

Deaths are the most common #1 priority, and they are also the most common highest-impact mistake: in 43% of reports, the first entry under Critical Mistakes is about dying.

The match data backs it up. Leaving Turbo aside, players died 9.4 times on average in the games they lost and 6.7 times in the games they won. 71% of losses had eight or more deaths, against 38% of wins.

What the reports describe is rarely a single bad fight. It is the same death repeated: staying in a fight after your damage is already spent, farming a lane with no vision on the enemy core, or walking alone to the same spot where you died five minutes earlier. A typical #1 priority in this group reads like "stop dying after you have used your spells", not "position better".

What to do: after every death, ask one question - did I have to be there? If the answer is no, that is the habit to fix. Count your deaths at the 20-minute mark in your next five games and try to bring the number down by two. Very few changes move win rate faster.

Mistake 2: Building for the Wrong Game

Itemization is the #1 priority in 27% of reports, and one item dominates the conversation. Black King Bar comes up in 87% of itemization sections, and in 67% of them the report points to it as missing, late or needed. The next most discussed items are Blink Dagger (57%), Aghanim's Scepter (56%), Linken's Sphere (40%), Lotus Orb (35%), Force Staff (31%) and Glimmer Cape (28%).

The pattern is not that players do not know these items. It is that they buy the same build every game regardless of the enemy draft. A core who goes greed items into three stuns, or a support who never buys a save item when the enemy has a Legion Commander or Bloodseeker, is building for an average game instead of this one.

What to do: at the 0:00 horn, look at the enemy draft and name the two heroes most likely to kill you. Then decide which item answers them - BKB against disables, Linken's or Lotus against a single targeted spell, Force Staff or Glimmer to save an ally - and plan when you will have it. If you play a core and the enemy lineup has three or more stuns, assume BKB is in your build until proven otherwise.

Mistake 3: Farming Too Slowly

Farm is the #1 priority in 26% of reports and the highest-impact mistake in 27%. In non-Turbo games, players averaged 554 GPM in wins and 434 GPM in losses. Part of that gap is simply that winning teams farm more, but the reports keep pointing to the same avoidable causes: idle time between waves, never stacking or clearing nearby camps, and moving to fights that were already over.

This is closely tied to the first mistake. Every death costs you the gold you lose and the farm you miss while you walk back, so a lot of "bad farm" is really "too many deaths" in disguise.

What to do: pick one simple rule and play five games with it: never stand still. When a wave is cleared, there should always be a next target - a camp, the next wave, a stack for later. You do not need perfect last-hitting to gain 100 GPM; you need fewer seconds spent doing nothing.

What Rarely Comes Up

The themes at the bottom of the table are just as interesting. Warding is the #1 priority in about 1% of reports and laning in about 2%. Teamfight participation is only 3%, even though nearly every report discusses fights somewhere.

That does not mean those skills are unimportant. It means that for most players most of the time, they are not the bottleneck. If you are spending your practice time on ward spots while dying ten times a game, you are working on the wrong problem.

How to Use This

Look at your last loss and check it against the three mistakes above, in order:

• Count your deaths. Eight or more is the clearest sign of where to start.
• Compare your items with the enemy draft. Did anything you bought answer their two most dangerous heroes?
• Check your GPM against your role. If it is low, look for idle time before you blame last-hitting.

Most games have one of these as the main problem. Fixing the right one is worth more than fixing three small things at once.

Find Out Which One Is Yours

These are averages across 584 games. Your own games will have their own pattern, and the fastest way to find it is to look at them one by one. Paste the Match ID of your last game into DotaMirror and the report will tell you which of these mistakes cost you that game, with the numbers to back it up.

Analyze your last game and see which of the three it is.