Courtesy of Flickr user Dave Newman under a Creative Commons license.

Much of Uber's appeal lies in how it qualitatively differs from traditional taxi service: the convenience of using an app-based dispatch, the predictability that comes from watching a car travel in real time to your front door, the ease of payment when you never have to pull out a credit card.

At the end of the day, though, what matters most for Uber's long-term success is whether it can offer not just a nicer but a cheaper alternative to other forms of transportation (whether that's a taxi or even a personal car). Because all of these other advantages will eventually be replicated by cabs.

In New York, where regulators have prevented the company from rolling out the true version of its budget service, Uber is apparently still not a bargain yet in many cases. According to a clever study from researchers at the University of Namur in Belgium and the University of Cambridge, a cab probably offers a cheaper ride than Uber in New York City for most of your trips under $35. Uber's price advantage only starts to kick in above that threshold.

The researchers looked at data on millions of taxi trips made in New York in 2013 that was publicly released last year after an open-records request. For every trip in that dataset β€” each includes origin, destination and fare data β€” the researchers queried Uber's API for how much the app would charge for the same trip on the cheapest service in New York City. Uber gives a range of possible prices, and the researchers took the average. Over the entire dataset, the median taxi ride was $1.40 cheaper. The full results are captured in the graph below, which shows the median Uber price for every yellow cab price:

There are a couple of big caveats here. The taxi data comes from 2013 and the Uber data from 2014. This study also wasn't able to take into account surge pricing during peak hours, a tactic cab companies aren't allowed to deploy in most cities. In reality, an Uber ride that's cheaper than a cab at 1 p.m. on a Wednesday may be more expensive at 10 p.m. on a Friday night.

A similar comparison would probably also look very different in just about any U.S. market other than New York. The researchers here used data on "UberX" in New York, but the service there doesn't mean what it does in D.C. The city still bans everyday drivers without chauffeur's licenses from deploying their personal cars as de facto cabs. So the people providing the "UberX" service in New York are still commercially insured and licensed black-car livery drivers. That means New York doesn't really have the cheapest version of Uber's service.

Although the graph above doesn't translate to other cities, a key argument behind this paper does: Consumers should have the same ability to comparison-shop prices for ground transportation that we've come to expect with air travel. In an ideal world for consumers, we'd be loyal to neither Uber nor Lyft, nor traditional taxi cabs β€” we'd exploit the differences between them, depending on the time of day, the destination and the length of the trip. The paper's authors are already working on an app to help consumers do that in New York.