A counter isnot a fact.It's a functionof time.
In pubs, "X counters Y" is a binary relation between two hero names. In professional play it takes arguments — position, lane, patch, item timing, teammates, the specific player — and returns different answers for different minutes of the same game. Every layer below is one of those arguments being varied. Drag, toggle and scrub each instrument; the point is to feel the shape, not memorize the list.
Most viewers grade a draft as a static comparison of hero strength. Professionals grade it as two overlapping curves and ask a different question: can we force the game to be decided inside our window? A counterpick that is only good from minute 12 to minute 25 is a real counterpick if your draft can also make minute 12 to 25 the decisive stretch — and worthless if it can't.
Below: seven layers, from the hero matchup outward to the metagame around the series. Hero names age out with every patch, so the instruments model mechanisms — the things that stay true when the numbers change.
Matchup
The narrowest layer, and the one everything else is built on top of. The mistake is treating it as a single relation. There are at least five different things people mean by "counter," they break in different ways, and almost all of them have an expiry date.
Five kinds of counter, and what dissolves each one
The category matters more than the outcome, because the mechanism tells you what the answer costs. A stat counter is a spending problem. A mechanic counter is a positioning problem. A win-condition counter can't be solved at all, only avoided.
The bar shows roughly how much of the counter survives contact with money and skill. Stat counters erode fastest — they're a head start, not a wall. Win-condition counters barely move, because no item makes your plan work again.
The same matchup, owned by different people at different minutes
Almost no counter is unconditional. It's gated on a level, an item, a cooldown, or on the enemy actually executing. Toggle the gates below and watch ownership of a single hero-versus-hero matchup flip back and forth across the game. The classic shape is three acts: they crush the lane, you own the middle, they win the sixth-slot game.
Turn every gate off and you get the pub model: one hero simply beats the other, forever. Every gate you add is one more thing a professional team is tracking in real time — and one more reason "who wins this matchup" is an incomplete question without a minute attached.
Winning the lane and winning the game are different jobs
This is the single most common confusion in matchup talk. A hero that dominates the first ten minutes may have no second act; a hero that loses the lane on purpose may be the reason the map opens up at minute 20. Teams routinely draft a losing lane, on purpose, because they've priced it.
The upper-left quadrant is where drafts are actually won: heroes that concede the lane and pay you back on the map. The lower-right quadrant is where pub win rates come from — and where professional teams happily donate the first ten minutes.
Draft
The draft is an argument about when the game will be decided, conducted with incomplete information, under a pick order that decides who has to commit first. Three sub-games run at once: hiding your own shape, forcing theirs into the open, and checking that your five heroes can actually finish a game.
Flex picks: buying ambiguity with picks you were making anyway
A hero that credibly plays three positions costs you nothing extra and hides your entire lane setup. The enemy has to prepare for every configuration your picks still permit. Pin each hero below and watch the number of lane setups they must prepare for collapse — that collapse is the actual currency being traded in the first phase of a draft.
The count is every assignment of five positions still consistent with what you've shown. Killing the enemy's ambiguity is a draft goal in itself — a ban that forces them onto a hero which can only be one position buys you more than a ban of a stronger hero that leaves three readings open.
Pick order economy: who has to commit, and who gets to answer
Every slot in a draft has a different information environment. Early picks are commitments made half-blind, so they get spent on flexible or universally strong heroes. Late picks are answers, so they get spent on the pins. Bans move in the opposite direction: early bans hit the format, late bans hit the specific plan you can now see forming.
Exact ban and pick counts change between patches, so this is the shape rather than a current rulebook. What doesn't change: information rises monotonically through the draft, and the value of a slot is set by how much of it you have. Bans also target people, not just heroes — a mediocre hero with one player's name on it gets removed all the time.
Answer coverage: drafting categories, not heroes
Professional drafters check off jobs. Not "is this hero good" but "who dispels, who saves, who clears summons, who breaks a passive, who takes high ground, who defends it." Build a five-hero draft below and watch the gaps light up. Most losing drafts are not weak — they're incomplete, or they're three heroes doing the same job.
Two failure modes are visible here. A gap is a job nobody does — it usually surfaces at a specific moment, like the first time you stand at their high ground. Redundancy is three initiators none of whom needs the others: strong on paper, and no better than one when the fight starts.
Setup and follow-up beat two independently good spells
A stun that guarantees an ultimate lands is worth more than the sum of the two abilities. Chains have a direction: something creates the condition, something else cashes it in. The draft question is whether your best abilities are reachable — a huge teamfight ultimate on a hero with no way to get into range is a card you never play.
Break one link and everything downstream loses value, which is why the enemy bans the setup rather than the payoff — the cheap hero that makes the expensive one work is usually the correct removal.
Lane
A lane is not a fight, it's a negotiation over where a line sits. Everything laners do — pulling, denying, aggroing, stacking, buying regen — is a way of moving that line, because the line decides who gets experience, who is safe from ganks, and how far the enemy has to walk to kill someone.
Creep equilibrium is the lane's actual state variable
Where the waves meet sets three things at once: whose experience range covers the fight, how close each side is to tower protection, and how long a rotating support needs to walk. Move the equilibrium below and watch all three change together. A lane that "feels" lost is often just a wave sitting in the wrong place.
Pulling is the strongest single lever here, which is why so much support skill is invisible: it happens in the jungle, thirty seconds before the thing it causes. Denies do double work — they remove enemy experience and drag the equilibrium toward you at the same time.
Losing a lane on purpose
Professional offlanes are frequently conceded before the game starts. The trade is explicit: the offlaner absorbs two heroes' attention while the position four is somewhere else entirely, and the accounting only works if that fourth hero converts the time into something — a kill on another lane, three stacked camps, uncontested vision, an early objective.
Lanes are scheduled, not improvised
Much of the laning phase runs on a script written before the game: who contests which rune, who leaves at what time, when the first stack goes up, when the wave gets cut. Which means a lane matchup can stop mattering the moment the script fires. The instrument below is a schedule, not a simulation — exact rune and camp timers move between patches, and the point is that a plan exists at all.
Notice how few of these events are reactions. A support who leaves at a fixed time isn't responding to the lane — the lane was drafted knowing they'd leave. Regen economy sits underneath all of it: who can afford to trade is decided by consumables bought before minute zero.
Timing
If you only take one layer from this page, take this one. Professional Dota is played against a set of clocks that are mostly public, mostly predictable, and mostly ignored by viewers. Fights don't break out because someone felt like it; they break out where windows overlap.
Item, level and objective clocks, stacked
Each row below is a window during which one side has something the other doesn't. Alone, each is a small edge. Where several overlap, the map compresses and a fight happens — the strip at the bottom counts the overlaps. Teams plan around sentences of exactly this form: "we have blink and they don't have their defensive item for another ninety seconds."
Level breakpoints are the ones amateurs miss. Talent tiers and a single point in an ultimate often move a fight more than the item slot everyone is watching, and they arrive silently — no shop animation, no sound, just a hero that is now allowed to win.
Most fights are a cooldown, not a matchup
A large share of professional aggression reduces to: their save is down for eleven seconds. Bait it, count, then commit. The instrument below is deliberately dumb — two cooldowns and a decision — because that's genuinely the whole idea, and it's the idea that separates a coordinated engage from five people walking forward.
This is also why teams throw away a support's life for what looks like nothing. Burning a save is a purchase, and the thing it buys is the next thirty seconds.
The late game is played against the respawn timer
After about minute thirty, the map state that matters is not net worth, it's who can come back. A kill without buyback is a long stretch of free map; the same kill with buyback available is a brief inconvenience. Every high-ground siege is really a bet on this number.
Cost and timer formulas shift between patches — the numbers here are a plausible model, not a lookup table. The shape is the durable part: buyback cost rises with wealth, so the richest hero is often the one who cannot afford to die twice.
Map & resources
Farm is not something that happens to a hero, it's something a team allocates. The map holds a roughly fixed amount of gold per minute, and a professional team decides in advance who gets which parts of it, then spends fights and vision defending that allocation.
Every zone belongs to someone
Click zones to assign them and watch the split. The total barely moves — what changes is whose curve gets steeper. This is why "our carry is behind" is often a routing failure rather than a skill one, and why a support who takes a camp on the way to a ward is making a real, if tiny, withdrawal.
In vision mode, notice that wards are not "map awareness" in general — each one answers a specific question. An offensive ward asks where their cores are farming; a defensive ward asks whether we can walk home. A deward that saves one support has paid for itself several times over in tempo.
Terrain does the work your abilities can't
The same team fight has a different answer in a choke than in the open, because area abilities are priced per hero hit. Drag the ability circle below at different terrain widths — the number of heroes you can catch is set mostly by the ground, not by your aim. Which is why teams fight at Roshan, on ramps, and in the trees, and refuse to fight in open lanes.
Uphill vision and miss chance stack on top of this, which is why the defending team's fight is nearly always better than their draft suggests — high ground is worth roughly a free item for the duration of the fight.
Split-pushing is an inequality, not a playstyle
It works when what you gain while they respond exceeds what they gain by responding, and when you can leave before the response lands. Both halves have to be true. The classic failure is a split-pusher who wins the arithmetic on damage and then dies to the answer, converting a small profit into a large loss.
Wave manipulation is the quiet version of the same idea: cutting a wave costs you nothing and forces the same response, and a catapult wave turns a lane into a siege resource that has to be answered on their side of the map.
Itemization
Items are how a hard counter gets renegotiated mid-game. Every threat has an answer, every answer has an answer, and each step costs a slot out of six. The arms race isn't infinite — it terminates when someone runs out of slots, and knowing whose slots run out first is a large part of what an experienced captain is doing.
The arms race, one slot at a time
Walk each chain forward. Notice that answers get more expensive and more specific as you go: the first response is a generic defensive item, the third is a niche purchase that does almost nothing against the rest of their team. That narrowing is the real cost — you're spending general-purpose gold on a single enemy hero.
Build order is also an information game. A visible early component announces your timing, so teams buy components in orders that hide the plan, or buy the obvious thing precisely to provoke a response they've already decided to punish.
The point where more damage stops being worth it
Defensive stats multiply rather than add: resistances stack against the same health pool, so every point of resist makes the previous points worth more. Damage does the same on the other side of the ledger. What you actually want to maximize is roughly the product — how long you live times how fast you kill. Slide the allocation and watch where the peak sits, then change the enemy's damage mix and watch it move.
This is why the correct answer to "they have a lot of magic damage" is rarely "buy more damage and kill them faster," and why armor items on a team are worth more than armor items on a hero — a defensive aura multiplies against every source of damage aimed at anyone standing in it.
Information & the series
The state of a Dota game includes things nobody can see: where the enemy support has not been for forty seconds, what your opponent thinks you'll draft next game, which hero got quietly stronger because an item changed. This layer is the one that separates good players from good teams.
Missing heroes are an expanding volume of danger
Each enemy you can't see occupies a region that grows with every second since they were last seen. Professionals move through the map by reasoning about these regions, not about the dots. Scrub the clock and drag your own hero — the warning fires when enough of those regions include the ground you're standing on.
Smoke breaks this model on purpose: it removes the growth of your own circles from the enemy's estimate while your team crosses the map. The counter isn't better vision, it's guessing which region they'd need to cross — and putting a sentry there.
What you show in game one is spent
A best-of-three is one long information game. A strategy revealed early gets answered later; a strategy saved may never get used. Teams deliberately underplay a pocket pick, or burn it in game one to bait bans that free something else. Hover the matrix to see how the payoff changes with the score line.
Format changes the answer completely. A single-game group stage rewards safe, forgiving drafts that don't need a plan; an elimination series rewards the pocket strategy and the willingness to spend it at the right moment. Patch reading is the same game against the developers — the heroes that got stronger are often the ones whose items or map changed, not the ones on the buff list.
What each layer varies
Every layer on this page is one argument to that function in the header being held still while another moves. Read the middle column as the question a professional is actually answering when they say a hero is good right now.
| Layer | The question underneath it | Failure mode |
|---|---|---|
| 01 Matchup | By what mechanism does this beat that, and when does the mechanism expire? | Treating a gated counter as permanent |
| 02 Draft | Can these five heroes finish a game, and have we hidden what they are? | Strong heroes, no plan for high ground |
| 03 Lane | Where does the wave sit, and what was the plan for when it stops mattering? | Playing the matchup instead of the script |
| 04 Timing | Whose windows overlap right now, and for how many more seconds? | Fighting on feel, off-window |
| 05 Map | Who owns which gold, and what terrain will the fight happen on? | Unallocated space, fights in the open |
| 06 Items | Who runs out of slots first in this arms race? | Answering a threat you were already beating |
| 07 Information | What don't we know, and what have we already shown them? | Spending a pocket strategy on a won game |
The through-line is the clock. Layers one, two and six all look like comparisons of strength and are really arguments about scheduling: every draft is a claim about which stretch of the game will be decisive, and most professional decisions are attempts to make that stretch arrive on time.