The odds of catching an unknown planet dimming a star are (pardon the pun) astronomical.

Even the New Horizons team needed a lot of effort to catch their new target passing in front of a star [1]. Granted, Planet Nine is a bigger target, but the odds are still pretty low.

[1] "NASA’s New Horizons Team Strikes Gold in Argentina" https://www.nasa.gov/feature/nasa-s-new-horizons-team-strike...

My intuition suggests that a planet far out and small would be very difficult to detect by its occlusion.

This page calculates that Pluto for instance spans only 3 pixels at the Hubble space telescopes resolution: http://www.planetary.org/blogs/emily-lakdawalla/2013/0214101...

If Hubble can barely see Pluto, how the heck did terrestrial telescopes find it in the first place?

To see something that acts like a point of light, size doesn't matter at all. Glowing objects are visible even when their area is a small fraction of a pixel. But your chances of noticing occlusion are much lower.

Wasn't the presence of a Pluto-like body _known_ because the orbits of other planets didn't match expectations based on the gravitational pulls of all the other known bodies in our solar system?

Not exactly. It's true that Neptune was discovered that way and that people found Pluto while looking for a "Planet X" that would explain other gravitational discrepancies:

In the 1840s, Urbain Le Verrier used Newtonian mechanics to predict the position of the then-undiscovered planet Neptune after analyzing perturbations in the orbit of Uranus. Subsequent observations of Neptune in the late 19th century led astronomers to speculate that Uranus's orbit was being disturbed by another planet besides Neptune.[1]

However it turned out that Pluto was not Planet X after all. After all Neptune and Uranus are respectively about 7900 and 6700 times more massive than Pluto.

In 1978, the discovery of Pluto's moon Charon allowed the measurement of Pluto's mass for the first time: roughly 0.2% that of Earth, and far too small to account for the discrepancies in the orbit of Uranus. Subsequent searches for an alternative Planet X, notably by Robert Sutton Harrington, failed. In 1992, Myles Standish used data from Voyager 2's flyby of Neptune in 1989, which had revised the estimates of Neptune's mass downward by 0.5%—an amount comparable to the mass of Mars—to recalculate its gravitational effect on Uranus. With the new figures added in, the discrepancies, and with them the need for a Planet X, vanished. Today, the majority of scientists agree that Planet X, as Lowell defined it, does not exist. Lowell had made a prediction of Planet X's orbit and position in 1915 that was fairly close to Pluto's actual orbit and its position at that time; Ernest W. Brown concluded soon after Pluto's discovery that this was a coincidence, a view still held today.[2]

[1] https://en.wikipedia.org/wiki/Pluto#Discovery [2] https://en.wikipedia.org/wiki/Pluto#Planet_X_disproved

The pixel size isn't what determines if occlusion happens.

The planet and the star can both be much smaller than a pixel, as long as the planet is big enough to cover the star.

Indeed. The article says that the hypothetical planet would be 2 pixels.

As other repliers have said, this sort of thing is very rare. Trans-Neptunian Objects are so far away that their visible size is absolutely tiny, so occulations are rare and last for maybe a second or so.

But! They do happen and are a great way to learn more about those objects as very briefly you can see something of them. The problem is that first you need to know the exact orbit, then you can predict occulations.

See https://en.m.wikipedia.org/wiki/(486958)_2014_MU69 for an example. It recently had some occulations that have helped better understand it.

If we pointed any of a handful of our most powerful telescopes at it we could image it directly. But how would we know it's planet nine versus something else? Some uncatalogued star or some undiscovered asteroid? We'd need to already know where it was supposed to be. Or, we'd need to have multiple images of the same section of the sky over a period of enough days to have the image of the planet move relative to the stars and be able to reveal itself as a distant planet.

The fact is, we simply do not image enough of the sky on a regular basis to have all this imagery just lying around. We do have some pretty good all sky surveys, but they don't use telescopes big enough to see something like planet nine. That's why it takes a dedicated effort to discover planet nine. It'll take imaging huge sections of the sky multiple times with enough delay between them to be able to see separation in the position of the object. Such large telescopes are very expensive and rare so their time is very valuable and in high demand.

Distant stellar transits will happen but we're not monitoring the entire sky 24/7 to the degree needed to spot all of them. If we were we'd be able to see the planet directly just as easily.

Our view of the sky is typically just a matter of seeing through little soda straws from one day to the next, so we miss a lot.

p9 is not believed to be aligned with the ecliptic. The most recent article I've seen estimates it to be inclined around 30° relative

[1] https://www.theguardian.com/science/2016/oct/19/planet-nine-...