Like all the reactors at Chernobyl, #4 was a graphite-moderated reactor of a type which has a "positive void coefficient"--PVC--which basically means that if the cooling water boils and a steam bubble forms the reactor will produce more power, rather than less. No reactor in America is built that way; by law, American reactors must have a negative void coefficient (NVC).
Using water as your moderator is fine, and with a NVC it ensures there's a negative feedback loop: if a steam bubble forms, the reactor produces less power, thus cooling and allowing the bubble to dissipate.
The problem with Chernobyl was that if the reactor produced too much power, and steam formed in it, the reaction could run away: the steam bubble meant more power, and more power means more heat means more steam, making the generation of even more power possible. And runaway reactions can rapidly go asymptotic.
The #4 reactor was rated to produce around 3 GW of thermal energy. Shortly before it blew up it made 320 GW thermal, over 100 times as much as its rated output.
Now--the fact is that with the safety systems engaged and with the reactor operated properly, the PVC issue can be avoided. Please note that there were at least three other reactors at Chernobyl which were operated safely until they reached their end of life and were decommissioned. The disaster at Chernobyl took place precisely because the safety systems were disabled and the reactor was deliberately mis-operated. It was done so as part of a test, but one that should have been done a lot earlier in the thing's operational lifetime.
Even so, Edward Teller--the father of the hydrogen bomb and one of my personal heroes--insisted that American nuclear regulations prohibit the construction of any nuclear reactor with a PVC, to prevent exactly the kind of event that happened at Chernobyl. There's absolutlely no reason to build commercial power reactors that way...except of course that it's cheaper.
Also, it's important to point out that American nuclear regulations require a containment building for any nuclear fission reactor. It's got to be encased in feet of steel and concrete so that if all the other safety systems fail and the core melts, no serious radioactivity can escape.
The Chernobyl reactor was contained in a sheetmetal building similar to a warehouse or light industrial space. Because it was cheap and easy to build it that way.
The point of all this exposition is so I can explain that we really don't--can't--know exactly what was going on in the #4 reactor when it blew up, or even shortly before. It wasn't designed to operate in that regime. There's no real predictability once a self-sustaining nuclear reaction gets to the knee of an asymptotic curve.
The reactor used enriched uranium for fuel; it had been operating long enough that the fuel was no longer pure U-238 and U-235, but a mixture of all kinds of elements and isotopes including various flavors of plutonium. The paper referenced by the above link makes a case for the possibility that some of the fuel rods underwent fission, as in "nuclear detonation" like an atomic bomb, only on a smaller scale, producing a yield equivalent to 75 tons of TNT.
It's difficult to produce a nuclear detonation that small, but perhaps not in the hellish conditions that existed in that reactor at that time. 320 GW is a lot of energy for a 3 GW reactor to produce; there's a lot of neutrons flying around--and it's just possible that there were enough neutrons to set off the kind of chain reaction that releases a lot of energy, really fast.
And then, after that, the steam explosion that blew the whole thing apart.
Some of the details which the paper says support the "small nuclear detonation" theory don't add up. For example, the "blue flash", that doesn't necessarily mean there was a detonation; I'd expect to see that sort of phenomenon from any nuclear reactor in the process of blowing its stack like that. Uranium reactors make neutron and beta radiation, and beta radiation is what makes that blue light. Overall, though, the paper's case is still pretty solid.
It doesn't really change anything, though. The Chernobyl event is still the worst nuclear accident in world history; it killed something like fifty people and rendered a large swath of countryside uninhabitable, at least according to the "no safe dose" theory. (The remaining radioactivity doesn't seem to bother the wildlife much, that's certain.) Even if a nuclear detonation was a component of the Chernobyl disaster, it still is not the indictment of nuclear power that nuclear power opponents wish it were.
And although they are now often mentioned in the same breath, Fukushima doesn't even come close to being the level of disaster that Chernobyl was. And Chernobyl itself--while very bad--was actually rather mild compared to some man-made disasters. Like Bhopal, India, where 2200 people were killed by a chemical leak.
But still, an interesting idea.