If you’ve played Dota at a high enough level, you’ve probably seen obvious account buyers. Questionable match history, high win rate, suspicious hero selection on the account and absolute cluelessness during the match. Nowadays, it is a lot rarer than it used to be, but each such encounter left me with a single, lingering thought: “what exactly are you trying to achieve here?”.
Winning a game as someone far below the lobby’s average is essentially out of the question. There isn’t a lot of fun in dying repeatedly either. The only pseudo-positive I can personally think of is that higher level lobbies expedite the learning process. Though, to be honest, there isn’t a whole lot of learning going on for the account buyer, in my opinion. Usually, the difference in the understanding of Dota is big enough to make learning any sort of lessons virtually impossible.
This is the fundamental issue with learning Dota in general: as it is an incredibly complex and deep game, there are so many moving parts to keep track and be aware of, that most humans will be overwhelmed if they try to take in all of it at the same time.
This is the reason stats websites like Dotabuff exist. They aggregate information from tons of various matches to give a quick overview of what we end up calling “the meta”. However, using this information well and actually improving through it is a learning process on its own.
One of the big, recent events in the Dota-world was that stats websites lost access to the highest level data. The information from the top 0.5% player lobbies is no longer available to us and no matter how you spin it, it is an information loss.
Whether this information was in any way relevant to the vast majority of the playerbase is a different question altogether. In my opinion, copying the top level builds and hero picks was not necessarily helpful when it came to learning Dota and sometimes even detrimental to the process.
When an average player takes a shortcut by copying a skill or item build from the professional scene, there is a high chance they inadvertently exhibit the same behaviour, as the aforementioned account buyer. They don’t fully understand why it works, and why it might be necessary for the players in the high level lobbies to make choices they make, they simply copy the builds, expecting positive outcomes.
Sometimes it works, reinforcing their belief in the strength of this approach. And when it doesn’t, there are always teammates to blame: after all, you are playing a top-level build that wins 55% of the game in 10k lobbies, surely you can’t be the problem in a 3k average game, right?
Here lies the problem: 1k average, 3k average, 5k average and 10k average players are playing fundamentally different games. As a broad stroke example: a 
Magnus in 1k should be preoccupied with simply executing a decent Blink + Reverse Polarity combo onto multiple targets. A 3k one can start learning the 
Harpoon + Skewer combo. At 5k they should be aware of instant hexes or silences on the enemy team and preemptively start using 
Black King Bar if necessary, before jumping in, since the enemies are likely to react in time or pre-cast their disables if Magnus is in vision.
In 10k+ this should all combine into a beautiful package where a Magnus can jump into one group of enemies, Horn Toss and Skewer them into another group of enemies for a massive RP, starting off with 
Shadow Blade off-screen to prevent pre-casting hex potential and using BKB during the Horn Toss animation on the enemy CC heroes, while they are still in the air, to maximize BKB effectiveness without getting disabled. Then do the same thing all over again after using 
Refresher Orb.
Striving to do that in 1k average lobby would be admirable, but there is a high chance executing this whole sequence is already going to be overwhelming to a new player. Moreover, there are going to be a lot of redundancies in terms of build as well: this Magnus in 1k doesn’t need Shadow Blade, since it is unlikely the enemy is going to pre-hex them or even have vision. They won’t need to press BKB before jumping in, as enemies are unlikely to react during the Reverse Polarity cast point. They probably won’t need a second round of spells, since the enemies are going to be clumped up regardless. And there are many, many more such small intricacies I am personally not aware of, since I am not a 10k Magnus player myself.
The main point still stands, though: simply copying a high level build in your average game will result in unnecessary, costly inefficiencies. And they will cost you games as a result. Looking at what works best at your particular level, understanding why it works and what the potential counterplays are, is what will make you a better, more knowledgeable player. When you start gaining MMR, using this knowledge to adapt to increasingly more difficult conditions will be much easier as well.
Copying builds is, by definition, a very rigid approach. There are definitely key items on certain heroes that are simply too good of a fit, but even they are not mandatory, if the a weird game you find yourself in is calling for something different.
Humans are also incredibly diverse and in a multi-faceted game of Dota, this means there are many different, successful ways of self-expression. Forcing yourself to copy specific playstyles is going to stifle your own progress, while potentially reducing your enjoyment of the game.
At the highest level of play the decision-space is certainly more limited, as some small build inefficiencies and high levels of greed are very punishable, but even then, there is room for creativity. Most of us will remember Topson building 
Diffusal Blade on 
Gyrocopter in TI Grand Finals and this sort of decision isn’t made by looking at statistics and copying the most popular builds, it is made by having a very solid, deep understanding of Dota.
Most of us don’t have time to sit and experiment for the whole day, but we all want to win and have our MMR go up. Statistics can help us identify trends and give a quick overview of what generally works in the current meta, but it can’t win you games. Improving as a player is the only way to get a higher rating. And improving as a player means analysing your own mistakes, being realistic about the level you are playing at and identifying what can be done to win more games in your particular circumstances; not by copying what works for players who are playing a very different game.
Copying “what works” isn’t the same as understanding “why it works”. That doesn’t mean you should go completely off-meta: each patch will definitely have its share of stronger than average heroes and items, but even when picking a “strong meta hero” make sure you are actively analysing why they are strong in your bracket and what limitations they might have if you find yourself in a higher one. You should not use statistics to find definitive answers. You should use statistics to ask the correct questions.
