Gas blowout could have been so sudden,driller could not catch it quickly enough. Then something had to ignite the gas. No smoking on drill floor. motors should be explosion proof. A tiny bubble of gas size of grain o rice deep in well expands to cubic foot at atmospheric pressure.
Well had been cased off and they were waiting for cement to cure before suspending the well and moving off.
Failure was sudden and catastrophic -- sudden explosion and fire 10 pm Tuesday night. Full details pending.
Cement obviously failed and allowed gas into wellbore.
This was an oil prospect not natural gas, but pay zone was below 18,200' (beaucoup deep) and pressured. If oil-based mud were in use, gas could have dissolved imperceptibly in the drilling-mud column (no warning volume-gain in pits as gas displaces drilling fluid), and come to surface w/o warning. This happened in the big Ixtoc blowout in 1978/9. Sneaky, fast-breaking, very deadly.
Alternatively, gas came to surface behind casing, evolving off oil reservoir and escaping to surface around failed cement job w/o observation until last minute. Kaboom.
Rams couldn't be closed (blowout preventers). Don't have the full story on that. With casing run, a new wellhead and BOP stack should have been in place.
Many available ignition sources: generators, light switches.
Rig also had 700,000 gallons of diesel fuel aboard.