blog, Jack's Stuff

Roadmap to 5K and Common Pitfalls

MMR’s just a number. But you can assign value to and make goals out of anything. For those of you looking to improve your solo queue ranked MMR, and get past what some consider a ‘4k trench’ (bogus in my opinion) into 5k territory, here’s a road map based on my own observations.

Emotional Management

Why do we start here first? Because I think emotional management is the single most critical skill for players to have when playing public DotA and looking to make the leap from 4 to 5K. Players are just good enough – and bad enough – that games can be swung based on morale, which influences teamwork and execution. The following is a pie chart distribution of ranked DotA 2 solo queue games.

That doesn’t seem right, you say. How come your team always gets the awful Peruvian trash who goes mid and feeds, or the immature rager who abandons, feeds couriers, or throws after first blood? As players we tend to remember these experiences more. We feel MUCH worse about losing an unwinnable game than better about winning an unlosable game. This is a combination of selective attention and negativity bias at work – and not the only time these factors will rear their ugly heads.

And it doesn’t even have to be random pubs who cause it. In this game, I had guanzo, gix, and Fogged on my team: but Merlini told everyone to ‘man up and random’ before the game. Much of Merlini’s team didn’t follow suit or picked heroes we couldn’t deal with. Guanzo, a 6.5k solo player, went 1-10 and started feeding intentionally before we lost rax to help facilitate the game’s end and an easy -32. In another game, I had Bamboe on my team. He promptly died five times in eight minutes and then abandoned, leading to another automatic loss. These games are going to happen, and they’re frustrating because the outcome is totally beyond your control. But that’s why you should detach from them – something totally beyond your control is not worth getting hung up over. And things often go the other way as well: in this game against Arteezy, I did little except hold down mid and avoid feeding, and my teammates carried me to an easy +26.

And here’s the best part, if you can get past the emotional bias: if you’re not part of the problem, if you never intentionally throw games or abandon them, the odds will actually be in your favor over time. This is simply because in any given DotA game, there are 9 players who are not you and are therefore unknown variables: 5 on the other team, 4 on yours. Therefore, if someone is bad or throwish enough to single-handedly derail their team’s chances of winning, and you’re never that person, this will tend to work in your favor.

Any team game where your fate is so closely tied to your teammates’ performance will be inevitably frustrating when your expectations are not met. But the right attitude, which may even minimize the number of ‘unwinnable’ games you have by not triggering throw-rage from your teammates, is to always try and make things better. Be supportive and positive, and try to get people to play together. Or take steps to curtail your own rage: my friend Dean routinely mutes himself or resolves to use nothing but the chat wheel in game, so that whatever frustration or anger he has doesn’t infect team morale. If you have don’t have something concrete or constructive to say, keep it to yourself. Goal-based communication helps tremendously.

And don’t just give up when things aren’t going your way – each win you get from a game you were ‘supposed’ to lose greatly swings your MMR. Here’s an example of one such game I was in:



Most importantly, be honest with yourself. Was the game really unwinnable, and exclusively your teammates’ fault? Is every single loss really because your teammates are bad, and couldn’t ‘be carried’? Does the other team really always play as a unit, and not make mistakes? Dunning Krueger is the bubonic plague of ranked DotA. You and every player in the game are making mistakes in each and every game, and you can always get better. Believing that you’re in a trench or that your teammates are always ‘uncarriable’ doesn’t help you improve, and raging just about never helps your team play better. Learn to separate your feelings from cold, careful analysis of your decisions and mistakes.

Emotional management alone is probably worth many hundred points. You just about never see pros or other very highly rated players get into tiffs with their teammates in pubs, and this isn’t a coincidence.


Hero Selection

This is the second major factor that can greatly swing your MMR. Do you enjoy just randoming last each game because you want to play a variety of heroes? Do you always just play one hero that you consider broken or strong? (Some of these exist in each meta, from yesterday’s Earth Spirits, Terrorblades and Broods to today’s Voids, Tinkers, Skywraths, Earth Spirits, and Broods). Do you always make sure to pick a ‘core’ who is going to get some farm somewhere, even if it’s the final pick and your team has only one support?

Hey, you can do whatever you want. But if your goal is to win more games, to have a bigger impact on games in that 73% gray area where your actions matter, you should be smarter about your picks.

In the 4K to 5K and above MMR range, hero picks and lineups do matter. Again, people are usually just good enough that serious handicaps in lineups can quickly decide the game. This doesn’t mean you need to wait down that All Pick timer to counter pick each time, or throw a fuss if your team cobbles together a ragtag lineup. It does mean that you should strive to avoid the following:

– picking heroes you don’t really know how to play, or play well. YOU ARE NOT ARTEEZY. Dunning Krueger again reigns supreme here, because everyone who has watched some pro’s stream thinks they can play like them too. I’ve seen way too many Tinkers or carry Nagas throw games and generate resentment from their team because they don’t have anywhere near the skill set to play like the pros, and ended up being huge liabilities.

– picking heroes who have already been counterpicked, or do not fit your lineup. There’s no faster way to lose a game than be part of the problem in picking a lineup that requires way too much experience and farm out of its heroes to work, and end up getting trashed in every lane. Again, you may feel like you’re sacrificing something for the sake of everyone else by making the ‘right’ pick, but if your goal is truly to win more games and lose less, this is the route you must take. This leads me to:

– picking a greedy jungler last when you’re about to start losing All Pick gold because you didn’t get a core sooner. I see this problem ALL the time; someone didn’t get a main laner and doesn’t want to be a second support, so they bust out the obligatory Furion or Doom so they can get core levels of farm. But they don’t play the hero well, and this puts incredible pressure on one support to do all of the warding/dewarding, still TP around to fights, and provide utility. This usually just snowballs into further resentment or that support’s back being broken, leading to a free loss.

– refusing to pick supports or believing that ‘I have to pick mid or carry for us to win’. This is the second debilitating instance of negativity bias/selective attention. People often perceive playing supports as an act of subordination: like, I’m humbling myself to support you, we better win. Under this mindset, you tend to remember losing games as a support more than you do winning them, because you EXPECT to win if you, the amazing player, have decided to make sacrifices to do so. You end up not remembering games where you lost largely because you played inadequately as a support, or games where you DIDN’T play support and lost because your performance as a carry or mid were not up to par.

But objectively speaking, it’s just not true that you’re lowering yourself by playing support (unless you’re awful at it). Support is a critical role, just like each of the others. If you happen to play support better than you do other roles, you’re still going to lose some games, but you add more value (and thus MMR) to your teams. Adding to this, in the 4k-5k range, solid supports are harder to find. Being able to fill this important niche alone can be worth several hundred points. As can making small sacrifices as a non-support to help fill team needs, like shelling out 150 gold to buy a direly needed set of wards. If your goal is to win games, you check your ego at the door and stop at nothing to be as useful to your team as possible.

This leads me to a final but more subtle psychological factor in hero selection: picking ‘broken’ heroes.

I’ve actually found that more often than not, when I was in the high 4000s making that final push, I tended to lose more when my teammates picked Void or Tinker. Why?

First, some players are so used to exploiting the imbalances on these heroes – imbalances which are more pronounced in lower MMRs – that they become overrated and don’t have the necessary skill and attention to detail for harder games. Second, picking broken heroes actually seems to often galvanize the other team into a commitment to destroy you. This is seen through either dedicated counter-picks (either a strong offlaner or duo offlane combo to shut down Void, or heroes like Venge/Shadow Demon/Disruptor, and heroes like Spectre or Clock for Tinker), or a general improvement in awareness and cohesiveness to crush your cheesy shit. The more a hero is perceived as broken or overpowered in a meta, the more better players will adapt and have counters ready, and the more determined they will be to beat it. If you’re not a good enough overall player, this will actually begin to hurt you as you rise in MMR. The people who once clumped up to die to Chronospheres or sent weak offlaners who couldn’t slow you down will now spread out and have ghost scepters/euls/swaps ready to save the one person you ult, and be sending strong lanes to keep you from farming.

The bottom line is, know what heroes you play well, and know which heroes add the most to your team in any lineup situation. Filling both requirements with your selection maximizes your chances of winning.


Closing Out Games

One of the major areas for improvement people in the 4k-5k range tend to have is maintaining leads and closing out games. Some of this goes under emotional management- players get a good start, seem to be rolling over the other team, and a few ill-advised tower dives, item choices or high ground sieges later, your lead has evaporated.  Like rubber-banding in a sports video game, there always seems to be a situation in each game where one team takes the lead, and the other surges back.

Hubris aside, the second reason for this was said indirectly by Sun Tzu: when you have people cornered, you leave them no choice. When a team’s area to defend is shrinking or consists of just high ground, they have no other option but to play together. Random pickoffs on greedy players dry up. Players get a sense of urgency, knowing that they have to defend their raxes at all costs. If the aggressor team isn’t careful in adapting to this, leads are very easily thrown away.

This is a frequent problem on the border of 4 to 5k, where there’s a frequent lack of patience or discipline. Keeping and converting incremental advantages into decisive ones is a skill easily worth several hundred points. Here are the most common game swinging mistakes I’ve seen/been a part of:

– forcing high ground without a major advantage / dead enemy heroes. also a part of this is not being content to slow siege or split push when abilities allow for it (like shrapnel or books)

– attempting risky Roshans

– buying back for no reason, or ignoring the value of enemy buybacks (or them not having it up)

– refusing to push when there is a golden opportunity, or when the clock is ticking on your lineup’s effectiveness

– getting a stupid item out of arrogance from a good start, believing that the good times will continue to roll. My friend recently played a huge hand in losing us a team ranked game by going Bloodstone after having a good start on Bristleback.

If you want to finish wins, there are clear ways to tighten the noose on opponents:

– restrict and deny vision. get a gem and clear out all enemy wards, blinding them and making it unsafe for them to leave base

– push out waves before preparing to high ground or Roshan

– smoke, bait, and pick off

– apply split pressure if you can’t take high directly

– wait for key items. Beyond obviously crucial items like BKBs, many pub teams don’t even have a clear Mek or Pipe carrier, when these items are almost always good or obligatory choices. I’d estimate that I see pipes in maybe 30% of games around the 4k – 5k zone…but just how often is Pipe a bad item? How often do you play against a team that is nearly bereft of magical damage? This also goes to communicating and observing your teammates. Seemingly just as common as teams with no Mek or Pipe are teams with more than one of each.

As with many concepts, most players know these things, but the devil’s in the details when it comes to getting better and crossing that 5k boundary. Which leads me to the final subsection:


Know and Grow Fundamentals

Even if your mechanics and ‘twitch’ skills aren’t elite, there are simply invaluable habits and skills you need to have if you want to progress beyond 4k. If you don’t know how to do these things, there is an ever-growing abundance of videos and guides that can quickly show you. Common areas for improvement include:

– lane management. For example, if you’re in the safe lane and have the support or hero matchups to zone out an enemy offlaner, it’s simply unacceptable for them to be getting gold or experience. Part of this is simply focusing on denying and positioning the wave – if you have to deny a bit more, attack a bit less, or even tank creeps before they get to your tower until they meet your next creep wave, do so. If you are a support, harass from positions that do not draw creep aggro, or abuse patches of fog/range to maximize your harassment. It is generally better to be ganking from the fog (either map fog or smoke), than be seen and accounted for on the map for the enemy to react to.

– pulling and stacking. Good supports maintain gold and experience income for themselves off stacks and pull throughs, denying enemy heroes growth in the process. But any hero can stack and pull by themselves, Helm of the Dominator or not. Get used to paying attention to the game time when you’re walking by neutrals. This goes into:

– minimizing downtime. The biggest thing that stands out to me when watching pros play is that their downtime is minimized. They’re not spending huge amounts of time milling around or doing nothing. They are either looking for kills, creating or farming stacks or waves, pressuring or taking towers, or otherwise gathering gold and experience. They don’t make as many return trips home, and they generally avoid pointlessly risky rotations. Here’s a situation: let’s say you’re playing a carry, and you just fended off a gank attempt from the other team, with enemy heroes either dead, limping home, or on the other side of the map. The super-safe play is to go home and heal. But the optimal play is to ferry yourself a salve, since you’re no longer under any immediate threat of dying, and continue reaping gold and experience out of your now vacated lane.

In a snowball game like DotA, every detail matters. If you’ve ever lost a teamfight that you were forced into but didn’t want because you were 100 gold short of BKB, that’s an obvious example for when being even the tiniest bit more efficient would have totally swung things the other way. Another Merlini example: he has enough map awareness and is so zeroed in on being efficient that you’ll often see him farming camps in the enemy jungle, with enemy heroes occupying the lanes above and below. You usually can’t get away with something like this until you’re a much better player, but efficiency should always be a goal.

– positioning and role. Much of this is obvious, or is something you should have internalized well before 5k: knowing where your hero fights effectively, focus fire and spell efficiency (not Laguna Blading a 40 hp hero solely to lock up the kill in a fight where it’s needed elsewhere/earlier) which spells to look out for and avoid, using the fog of war to obscure your ganks or escape/complicate theirs. All of these can be improved on. But when I say role, here’s something big that I see being an issue in these games all the time, especially for heroes with big teamfight ultimates or skills (Ravage, Reverse Polarity, Black Hole and the like): Are you better off initiating or counter-initiating? I’ve seen this big-picture decision single-handedly swing games.

If you’re going to be initiating, it’s usually because your team has enough followup in terms of damage (or more big ults) to decide a fight with a clean jump, or you’re killing one hero that means everything to the other team. But for most games in this range, people will have some inclination to spread out or otherwise be aware of the big ults. You often can’t get good initiations without great split-second decision making, or group mistakes. If you’ve seen games where there isn’t enough followup after a Ravage, or an aggressive Black Hole or Chronosphere gets immediately countered or stolen, you know what I’m talking about. The simple truth is that sometimes big ults, especially if your team doesn’t have more than one, are best used as counter-initiation, once the enemy is distracted or groups together to fight or pursue. You should generally use Chronosphere liberally early and mid-game for single pickoffs which accelerate your farm and experience, but you can’t always be opening with it later on because it shows your hand and commits you. A big spell in reserve is one which must be accounted for and feared.

– flanking/engagement angles, terrain utilization. This kind of goes under positioning, but flanking is a tactic I don’t see nearly enough in these games. One of my favorite heroes is Queen of Pain, and what I like to do in enemy deathball situations is to find angles of approach from the fog or elevation. The benefits, more safely accomplished with a defensive item such as Linken’s Sphere or Black King Bar, are numerous: it’s easier to land big aoe ults like Sonic Wave or anything else if you’re coming from a place the opponent does not expect, and it puts pressure on the other team. Do they peel off, split up and try to burst you down? Are you hitting them at a moment when the front line is going on the rest of your team, effectively dividing them? Are you getting heroes out of position for your own team’s initiation? Are you able to deal more damage – once the game progresses Queen of Pain usually can’t kill as effectively by going all-out, and needs several rounds of ducking and weaving to kill stuff – because you can’t be locked down and handled without serious coordination from the other team? Anything that forces the other team into tougher decisions or exploits their lack of information / cohesion becomes doubly valuable in public games.

– carry TPs and know when and how the team needs you. There are few things more frustrating than realizing that your team isn’t going to win because you’ll never get as many heroes to the point of conflict as the enemy, because someone is prioritizing a small amount of their own gold or experience over helping. Don’t be part of this problem, and don’t let the desire for a few hundred extra gold put you or your team in a horrible position. At all points in the game, don’t ignore runes if you can help it. Runes can snowball into a significant advantage or decide how the midlane goes – same with rotations and ganks. These things all add up, and given how much players like to scrutinize how their mid performed, they all matter. If you’re playing a support and not helping your mid control runes or launch/counter ganks, don’t be surprised if he struggles as the game goes on. And sometimes, regardless of what hero or position you’re playing, dying for the team is the best way to ensure that team fights are won. Even pros frequently make this mistake, as it stems from sometimes overvaluing themselves or their hero: productive deaths are a thing. ‘Space created’ can mean anything from playing aggressively in an offlane to make sure enemy supports must stay honest to zone you, to absorbing damage and focus fire in team fights to ensure that the enemy team gets out of position.

Conversely, do not be afraid to give up objectives or avoid bad/lost fights in order to get something in return. I watched a Merlini phantom lancer game where he and Fogged came back from a very early (sub 25 minute) mega creep deficit to win. They remained calm about getting behind, didn’t feed in defending raxes against a Lycan they couldn’t fight, and slowly edged their way to an eventual throning. Winning a game like that is not only hugely satisfying, but a 50 MMR swing. And swing games is what increases your MMR. 95% or more of players would have just given up. But they saw solutions and fought for them, and eventually snatched away a win.



As you play more and more ranked games and variance runs its course and your sample size gets big enough, your MMR will be more and more accurate. The only trench is one that you put yourself into – continue looking for ways to improve, and your rating will too.



Leave a Reply