I’ve recently gained quite an appreciation for Crazy Taxi, having only played it a handful of times as a kid. That’s not to say that I’ve been playing any crazy amount of it; I’ve put some effort into learning some of the game’s “Crazy Moves,” but it’s a disc that I’ve been popping-in only when the mood strikes me. While I love the game’s iconic aesthetic qualities and snappy controls, it’s been its unique structure that has inspired a game design that I’ve been working on. Specifically, although I’m not good enough at the game to speak for its execution, I think the core of Crazy Taxi’s scoring system beautifully enforces this structure.

If you’re unfamiliar with the game, the core gameplay involves picking-up passengers from predefined locations and taking them to their desired destinations as quickly as possible, with the car needing to have reached a full stop in each instance. There’s a global timer, ending a run when it reaches zero, and each collected customer has their own timer which is added to the global one. Reaching a customer’s destination on-time converts their remaining time into a bonus that’s added to the global timer, with their payment added to your earnings. Alongside an increased fare for speedy delivery, you earn tips by giving passengers some excitement via jumps, drifting and near-misses, with a combo multiplier that’s ended by colliding with traffic. If a customer’s timer reaches zero, they’ll get out of the taxi and you’ll receive only the devastation of wasted time.

So, the general loop of a Crazy Taxi run is collecting customers, delivering them to their destinations quickly, and thus getting more time to get more customers, all to earn more money. The large base fares are incentive enough for you to deliver as many passengers as possible, and the combined bonuses to both earnings and the timer make it clear that speedy journeys are the key to success. As with driving real, less crazy taxis, route optimization plays a big role in chopping-up travel time: you can get the gist of your destination’s direction from the arrow atop the screen, but will quickly outsmart it by remembering locations and finding shortcuts, simply by gaining experience of the area. Thence, the game actively rewards knowledge with score potential, largely separate from mechanical execution.

In the pursuit of tips, making journeys fun for passengers involves executing advanced manoeuvres, ideally without slipping-up (so as not to break combos). Narrowly avoiding traffic, drifting and jumping at every opportunity—plus optimally accelerating and decelerating with precise faffing of the gear stick and pedals—takes a good deal of practice. Something as mundane as the direction in which the taxi’s facing when it comes to a halt can determine which customers are reachable before the timer runs out, and gradually gaining full control of the car is incredibly satisfying, much like improving at a fighting game. Not only is it intrinsically fun to drive with style, but the tip system means it’s extrinsically rewarding. Importantly, none of this bonus score means anything if you don’t deliver on time, as the fares and tips are only paid on arrival; you have to be a good taxi driver before being a crazy one matters.

Not only does Crazy Taxi’s score system separately reward mastery in knowledge and execution, but it does so in a manner that contextually makes sense: the game does track your number of customers, but the score that counts on the leaderboard is your total earnings number. Getting more customers is merely an avenue to more fares and more opportunities for tips. Furthermore, it’s clearly desirable to be the taxi driver with the highest earnings in the area, for both the player and the character they’re controlling; the score is diegetic. In this way, I feel that the game’s scoring really works to emphasise the core objective and intended loop, with the player’s goals entirely overlapping the playable characters’. The interface is also quite clear about when you’re earning fares and tips, when they’re getting paid-out and the degree of urgency at any one moment with respect to time, such that your objective remains obvious.

As a game limiter, the timer is an overt tether to the game’s arcade format, and isn’t exactly sensical in-universe: while the people in Crazy Taxi are either bafflingly busy or absurdly impatient (perhaps akin to the people in your area ;¬) ), there’s little within the game to suggest that any of the playable characters need to finish their shift within the next few minutes. Functionally, however, I’d say it’s a great way to encourage players to work on their fundamentals as taxi drivers. With initial attempts almost guaranteed to be short, players can immediately gain familiarity with the streets and customers within the starting area, picking-up and delivering passengers even slightly more quickly, and seeing gradual increases in those first fares. Guaranteeing that newer players are spending more time in the starting zone also allows them to improve at driving, with those wider initial streets providing ample space to test the limits of each car’s handling. Plus, the early zone’s light traffic and steep hills improve the likelihood of players noticing that tips are rewarded for near-misses and jumps. Interestingly, my niece and nephew seemed excited about the prospect of earning any money at all, let alone getting higher amounts in successive attempts, assumedly since the base fares already seem like a lot of money for the time they take.

As a point of comparison, I love (most of) the Devil May Cry games, but I can see their more abstract scoring systems leaving newer players wondering as to how they can improve. Each game’s loudest UI element is its Style Rank Gauge, which is clearly somehow measuring players’ “style,” but it might take some time (or some reading of the manual) for a player to realise that it’s bad to be frequently reusing attacks. An acceptable assumption would be that any technique that kills enemies quickly is best, but that’s often not the case. Admittedly, the confusion could be exacerbated because the game has two measures of success, being score and continuous plot progression, and many players who care about one won’t care about the other at all. Particular player behaviour can be encouraged in one measure, while being actively discouraged in the other, such as how using healing items to stay alive means not receiving an important “No Items” score bonus at the end of the level. Critically, while there are numerous contributing factors, Crazy Taxi has one core goal of making as much money as possible; improving at the game means finding ways in which to make more money.

Considering Crazy Taxi’s strictly run-based format, perhaps this comparison is unfair, but I chose to talk about Devil May Cry because of its visual emphasis on its scoring system, and how it overtly rewards stylish mastery of execution and knowledge. Improving scores and rankings in DMC games is a lot of fun, but while the style meter does sell the importance of playing stylishly, it doesn’t necessarily encourage intended play. I’d say that this is from the Style Rank’s degree of abstraction—in objectively measuring subjective quality—and the delay between playing the game and seeing the actual score awarded. In the first Devil May Cry, the Style Rank also only has an indirect impact on players’ scores: the game only measures time and Red Orbs obtained, and the latter are collected within a short time by achieving a high Style Rank. For me, the real elegance of Crazy Taxi’s scoring, via its approach to tiered rewards and its diegetic quality, is its immediacy in getting you to understand your objectives and intended behaviour.

Originally written on 21/7/24, after having got my brother to play Crazy Taxi with his kids. Edited for sharing on 25/7/24, and again on 27/12/24, before being tweaked and posted on 3/1/25.