One patch. Three different ways to improve.
A returner rebuilds the decision map, a limited-match learner trains one repeatable question, and a grinder searches for replay counterfactuals.
Three matches to understand the value of a minute again
Separate 7.41e economy changes from hero balance that does not affect your decisions.
Record the first key item minute and what the team achieved after it.
Before a late objective, compare the component with buyback instead of auto-buying.
If role or hero changes, do not carry the previous item timing over as a universal benchmark.
Portraits mark a role lens, not a hero tier or build recommendation.Learn through one objective, not the entire replay
Keep one immediate reason to save gold: Roshan, a tower or high-ground defense.
Enter net worth, current gold and seconds to objective before opening the shop.
Grade the spent gold after the next fight: it worked, it failed or no fight happened.
If the team cannot re-enter after buyback, the reserve loses value. Include position and TP access.
Portraits mark a role lens, not a hero tier or build recommendation.Review the economy as a chain of forks
Capture four points instead of one final GPM: purchase, objective, death and buyback.
Test what changes with a two-minute earlier purchase or a preserved reserve.
Record the condition where greed flips from positive value to lost tempo.
If no realistic input changes the answer, the model may be measuring the wrong decision.
Portraits mark a role lens, not a hero tier or build recommendation.The same 1,000 gold has a different job
A carry buys throughput and timing, an initiator buys access and reliability, and a support buys information and preserved team value. This is a conditional framework, not one tier list.
The next slot must either accelerate the map or win a timing
Primary question: Is there enough safe farm to repay greed before the mandatory fight?
Accelerator
The purchase needs enough waves/camps and a real window before the next pressure point.
Timing converter
Immediate power dominates growth when one fight opens Roshan or high ground.
Liquidity
A reserve matters only with a TP or positional route back into the fight.
Anti-Mage is a visual example of the carry lens, not a hero tier or universal build order.
Test the gold decision →
The first purchase creates access; later gold improves execution
Primary question: Does the item create a new way to start or survive the fight?
Access
Blink may change the action set rather than damage. That is a discrete economic payoff.
Reliability
The next gold raises the probability that purchased initiation is actually realized.
Team monetization
A cheap consumable can realize an expensive timing faster than another component.
Axe marks the initiator archetype. Hero, draft and cooldowns always override a static order.
Test the gold decision →
A small budget should buy probability of a team outcome
Primary question: Which spend creates value without perfect personal net worth?
Information
Vision lowers the chance of a bad fight and improves the next objective decision.
Value preserved
Utility pays through a saved core, position or cooldown rather than personal GPM.
Cheap catalyst
Smoke turns information and team timing into an attempt to secure a kill or objective.
Crystal Maiden is a visual marker for the support lens. This is a capital framework, not a required build.
Test the gold decision →