So I see Dota team as a combination of 5 players, for which every has a combinaton of two most general skills (macro and micro), for which they can make two most general types of errors (overstep and understep).
Games are lost because certain overstepping (overestimation of enemy's macro limitations: thinking they can't rax you, can't kill you all in 5x5, can't take roshan, etc. or overestimation of enemy's micro limitations: thinking you can 1x1 a certain hero, thinking their Meepo wouldn't be able to play properly, thinking you could gankm when you couldn't, etc.) or understepping (underestimation of enemy's macro/micro the same way, just in reverse).
Dota is all about knowing the best what you can and can't do without getting punished at a certain point in time, as a certain hero and what would be the best thing to do out of all available actions.
It literally is what people call skill.
A hypothetical perfectly bad player would be a player, who overestimates/underestimates the outcome of his every macro/micro decision.
A hypothetically perfectly good player would be a player, who doesn't overestimate/underestimate anything at all, but has an ideal vision of what to do at any moment as any hero, individually or as a team, always spot on with everything.
hold up
so what you're saying is...
let me rev this
you're saying
so you're really saying
that the BETTER TEAM wins?
not so sure about that
Some time ago I was thinking something similar to your macro/micro typology (without any particular analytical purpose), only with finer distinction between the categories:
1) Micro level would include the above mentioned immediately distinguishable attributes which you can observe directly on the player's screen - reaction time, lane positioning, landing a hook, dodging a spell, the ability to predict your opponent and lasthit/deny accordingly etc. You don't necessarily need to actively think about why you are performing these actions and how they fit into the larger plan. Instead, you need a firm and precise hand.
2) Indermediate level means knowing how to control your hero on a larger scale - how to position yourself before and during a teamfight, when to rotate and gank, whether you should hex their wyvern or shaker before the fight begins and so on. The main difference is that it is not "physically" difficult to perform these actions - it requires the same amount of effort to click uphill or downhill - but rather to decide which of these options will give you an advantage.
3) Macro level would then be a mixture of what is called "meta" and large scale decisions of your team. Ok, you've got your carry, you are approaching a teamfight...buy should you be there in the first place? Wouldn't it be better to buy smoke, split your team, make a distraction and rat their racks in the meantime? Was it smart to choose deso Chen as your carry in the first place? These questions often exceed the scope of one particular game and you have to think about them in terms of what you've seen in the past few weeks.
Or perhaps it's useless bullshit, the fuck do I know.
Lütfen yorum yapmak için giriş yapınız
In this thread I am gonna do what I usually hate seeing myself.
I am going to try to theorycraft the "win" scenario in Dota in the most abstract way possible the way I see it.
You have 5 players on each team, their final goal is to destroy the ancient of the opposing team.
Before destroying the ancient, they need to get an advantage to do that (successfully take preceding objectives/accumulate financial advantage/etc.)
In order to do that, each player has two inner core functions (equations) (functions of multiple other subfunctions): understanding of their macro decision limitations and understanding of their micro decision limitations, subfunctions for these two core functions may be reaction speed, morale loss or gain, experience, other human-specific factors. In the end, these two would form some kind of a differential equations system.
And also one general processing function (GPF) to process the resulting output of the forementioned functions and return the final victory probability.
There would be two general types of errors: overstepping error (overestimating your abilities) and understepping error (underestimating your abilities).
Macro decision limitations function would thus define a specific player's ability to understand what they could do on the map as a team and what they couldn't, in other words, the likelihood of a specific's player decision result returning 1 or 0.
Micro decision limitations function would define a specific player's ability to understand what they could do with their specific hero (know where they can kill and get away with it, know where they could farm and get away with it, harass, deny, destroy objectives as a solo hero, etc.), returning 1 or 0.
Decision would be defined as any shortest action a player can perform, permitted by the game engine.
In the end, the likelihood of winning a game by a team is given by the GPF output, which would, I assume, be some kind of function of macro+micro outputs sum, with different weight coefficients.
This is a very broad, not detailed, wireframe vision that I have.
In reality to formally test such a model, one would need a deep insight in random processes/probability theory/neural networks/discrete math, which, unfortunately, I lack.