Street Drop · Food for Thought
This game re-specs the location-guessing genre by replacing street imagery with a distilled set of environmental clues that narrow where a place sits on the map.
The location-guessing genre asks a player to deduce where a place is on a world map from a set of cues, then mark that place with a pin. Each round presents a location, and the player places a pin to indicate a guess. The score depends on how close the pin sits to the actual location, with more points awarded for closer guesses and a maximum score for a perfect match. This structure creates a loop of observation, hypothesis, and feedback that has defined a broad category of geography games. The genre emerged in May 2013, when the Swedish IT consultant Anton Wallén released a browser-based game that gave the entire category its name. That game, GeoGuessr, runs five rounds worth up to 5,000 points apiece, and established a pattern where players examine cues and locate a spot on a map, with scoring based on proximity to the true location.1
Street Drop enters this genre as a kin game, preserving the core loop of deducing a place and dropping a pin while changing what cues are available. Instead of street-level photographs or panoramic scenes, each round hands the player a short written dossier of generalised environment clues under the heading “What you can see.” The clues give the side of the road traffic drives on, the hemisphere, the climate, the vegetation, the language on signage, the architecture, and the currency — in that order. The player reads them, clicks a flat world map to place a pin, and drops it to lock the answer in. The closer the pin, the higher the score. The mechanism remains the same as the genre's foundation: a place must be deduced from cues and located on a map, with scoring by proximity to the true spot. The game's own curated dataset of sixty-one cities provides the clues that guide this deduction.2
What distinguishes this re-spec is the abstraction of the clues. Rather than presenting visual details that require interpretation of shadows, foliage, or signage scripts, the game provides a structured set of fields that summarise environmental characteristics. One field names the language appearing on signs. Another says whether traffic drives on the left or the right. A third places the location in the northern or southern hemisphere. Four more give the climate, the vegetation, the architecture, and the currency. These seven short phrases are everything the game knows about a place, and the player must synthesise them into a geographic hypothesis.2
Each round begins by choosing a target from the game's curated dataset: sixty-one real cities spread across fifty countries, each stored with an approximate latitude and longitude and its set of generalised environment clues. The choice is deterministic. A small seeded generator turns one number into one index into that list, so the same seed always yields the same city. In the daily mode that seed is a hash of the calendar date in UTC, which makes the day's drop identical for every player; the practice mode seeds itself from the clock instead and advances the seed on every new drop, so a practice round is nobody else's round.2
Two difficulty settings decide how much of that dossier the player sees and whether a clock runs. The setting the game labels Classic reveals all seven clues with no timer. The setting labelled Speed reveals only the first four — driving side, hemisphere, climate, and vegetation — and gives the player thirty seconds; the signage language, the architecture, and the currency are withheld altogether. Speed compensates by multiplying the final score by 1.25. Neither setting changes the core mechanic of placing a pin. What is easy to assume and wrong is that the clues trickle out as the round runs: every clue the setting allows is printed at once the moment the round starts, so the fixed clue order decides which clues Speed keeps and the sequence they are listed in, not the pace at which they arrive.2
Once the player drops the pin, the game computes the great-circle distance between guess and target with the haversine formula, on a sphere of radius 6,371 kilometres. That distance is converted into points by an exponential decay: five thousand, multiplied by e raised to minus the distance divided by two thousand kilometres, rounded to a whole number. A perfect pin in Classic therefore scores 5,000; in Speed the 1.25 multiplier carries a perfect pin to 6,250. Either way the outcome is graded rather than binary, and the game reports both halves of it — the distance in kilometres and the score.2
The seven clue fields work as filters that narrow the set of plausible locations, but they are nowhere near equally sharp. The signage language is the strongest of them: forty different signage descriptions are spread across the sixty-one entries, and thirty-one of those appear exactly once, so a clue reading Icelandic, Thai, or Amharic identifies the city outright. The driving side looks as though it ought to halve the world, and it does not. Nineteen of the sixty-one cities drive on the left and forty-two on the right, so being told that traffic keeps right — the commoner answer — removes only nineteen candidates, under a third of the field. It is also the clue listed first, the one the player has to start from.2
The hemisphere clue divides the planet at the equator, and it is lopsided in the same direction. Forty-nine of the sixty-one cities sit north of it and only twelve south of it, so “northern hemisphere” eliminates about a fifth of the field while “southern hemisphere” eliminates about four fifths and is by far the more valuable answer to draw. The climate, vegetation, and architecture clues add specificity by describing conditions that cluster in latitudinal and regional bands: tropical climates concentrate near the equator while continental and subarctic ones occupy higher latitudes, vegetation follows climate, and architectural styles reflect regional building traditions. The currency clue ties the location to a national monetary system, though it is blunter than it sounds — the euro alone covers ten of the sixty-one cities.2
The clues matter most in combination. A player told that traffic drives on the left and that the location lies in the southern hemisphere is already down to seven of the sixty-one cities, and climate and vegetation will usually separate those seven. In the Classic setting the signage language and the currency then close the gap almost completely. In Speed neither is on the page, which is why the fast setting is not simply a shorter version of the slow one — it asks the player to finish on climate and vegetation alone.2
The player's task is to translate abstract clues into concrete coordinates. Each target's latitude and longitude are approximate but real, fixed when the city was added to the dataset. The player must estimate where those coordinates lie from the environmental profile alone, which means knowing how geographic characteristics distribute across the planet. Languages cluster in particular regions. Driving sides follow historical and colonial patterns. Climates form latitudinal bands. Vegetation mirrors climate zones. Currency systems are tied to national boundaries. Architecture reflects regional traditions and available materials. The pin itself goes down with a mouse click on an equirectangular world map, where longitude maps straight across and latitude straight down, so the click becomes a latitude and longitude by two divisions and nothing more.2
It is worth being precise about what the seed actually controls, because it is less than the word “deterministic” suggests. The seed picks the target city and nothing else. Which clues appear is decided by the difficulty setting, not the seed, and the order they are listed in is a fixed constant compiled into the engine, identical in every round ever played. The shared-experience guarantee holds only for the daily drop, whose seed is a hash of the UTC date; practice seeds come from the clock, so two players are most unlikely to meet the same practice city. What the determinism does buy is reproducibility: a given seed and difficulty will always rebuild exactly the same round, which is what makes the engine testable.2
The scoring geometry rests on the great-circle distance between guess and target: the shortest path between two points on a sphere, measured along the arc of the great circle through them. The haversine formula is the standard way to compute it from latitudes and longitudes, and is used here because it stays numerically well-conditioned at small distances, where the older spherical law of cosines loses precision. Treating the Earth as a sphere rather than the flattened ellipsoid it really is costs some accuracy, but little: great-circle formulas applied to longitude and geodetic latitude are accurate to within about half a percent, and the engine uses the conventional mean radius of 6,371 kilometres.3
The dataset is the foundation of every round: sixty-one real cities across fifty countries, each with an approximate latitude and longitude and its seven generalised clues. It runs from Reykjavik at sixty-four degrees north to Melbourne at nearly thirty-eight degrees south. It also holds pairs close enough that mixing them up barely hurts — Kyoto sits 363 kilometres from Tokyo, so guessing the wrong one of those two still returns 4,170 points of a possible 5,000, and Melbourne is 713 kilometres from Sydney, worth 3,500. Each entry is a real place with real environmental characteristics, compressed into seven short phrases.2
The generalisation of environmental data into these seven fields is a design choice that shapes how players interact with geographic information. Rather than presenting detailed street-level imagery or complex environmental data, the game provides a structured summary that players must synthesise. This abstraction requires players to think about how geographic characteristics distribute across the planet and how those characteristics constrain the set of plausible locations.
The clue order is a fixed constant, and it does two jobs rather than one. It sets the sequence in which the clues are listed on the page, and — because the Speed setting simply takes the first four of them — it decides which three clues Speed throws away. That ordering is a real design judgement. Putting driving side and hemisphere at the front means the fast setting keeps the two coarsest filters in the game and discards the three sharpest: the language, the architecture, and the currency. A Speed player is not solving a smaller version of the Classic puzzle. They are solving a blurrier one, for 1.25 times the points.2
The scoring rule is one line of arithmetic. The haversine distance in kilometres goes in; five thousand, multiplied by e raised to minus that distance divided by two thousand, comes out, multiplied by the difficulty bonus and then rounded to a whole number. Two thousand kilometres is the whole character of the curve: every further two thousand kilometres of error multiplies what is left of the score by about 0.37. A pin two hundred kilometres off keeps 4,524 points. One thousand kilometres off keeps 3,033. Two thousand kilometres off — the width of a continent — still keeps 1,839.2
That shape is forgiving where the genre needs it to be and unforgiving where it does not. Because the decay runs on absolute distance rather than on the ratio between two distances, what matters is how many kilometres you missed by, not how many times over: doubling a small error and doubling a large one cost quite different amounts. The first two hundred kilometres cost under a tenth of the maximum, so a player who identifies the right country almost always banks most of the points. The tail, though, falls fast enough that a guess on the wrong side of the world is worth nothing worth having — ten thousand kilometres out returns thirty-four points, and only past about eighteen thousand four hundred kilometres, close to the antipodal limit of roughly twenty thousand, does the rounded score finally reach zero.2
The evaluation step is the bridge between geographic measurement and game scoring: it takes the round and the guessed coordinates, measures the arc, applies the decay and the difficulty bonus, and hands back both numbers for display. One edge of it is worth noting, because it is the case the formula cannot speak to. If the Speed timer runs out with no pin on the map, there is no arc to measure at all, and the round settles at zero. A guess and the absence of a guess are different things, and only the first of them has a distance.2
The location-guessing genre persists because it transforms geographic knowledge into an interactive challenge. Players must observe cues, form hypotheses, and test those hypotheses against a scoring mechanism. The genre's structure creates a compelling loop of deduction and feedback. Each round presents a new puzzle, and each guess provides information about how well the player understood the clues. The scoring mechanism rewards accuracy and provides a clear measure of performance.1
Street Drop's re-spec of this genre preserves the core loop while changing the nature of the cues. By presenting a curated set of generalised environment clues instead of street-level imagery, the game shifts the focus from visual interpretation to geographic reasoning. Players must think about how languages, driving sides, climates, and other characteristics distribute across the planet. They must synthesise multiple clues into a coherent geographic hypothesis and translate that hypothesis into coordinates.2
The daily seed and the curated dataset give that reasoning a consistent frame: everyone gets the same city on the same date, and any given seed rebuilds the same round exactly. The scoring rule turns the result into a clear measure of geographic accuracy. The genre's foundation, deducing a location and guessing it on a map, with scoring by proximity, remains intact. Street Drop enters this genre as a kin game, contributing to the broader category of location-guessing games that ask players to locate places from cues and score by how close their guesses come to the truth.1