It's quite clear the middle two are different colors, and reading left to right, #3 is darker than #2. #4 looks to have some remnants of shininess, and so appears to be a gold tolerance band. There's another nearby green resistor with a gold band as well.
So from left to right, my guess is:
Dirty white (or maybe yellow?)
Red
Brown
Gold
So that's, uh, erm...
9
2
x10Ω
-----
920Ω +-5%
This resistor calculator converts the ohm value and tolerance based on resistor color codes and determines the resistances of resistors in parallel or series.
www.calculator.net
I have no idea whether that actually makes sense or is an actual component one can buy, mind you.
But #3 could be an overheated orange, perhaps? Basically, #2 and #3 could be red, orange, brown, or maybe even violet depending on what you think the effects of overheating and age might have been.
Personally, I'd try to wipe off some of the grunge with WD-40 or alcohol or similar and get some good strong incandescent (full spectrum; not LED) lighting in there. If that doesn't bring any enlightenment, maybe try scraping or lightly polishing or sanding part of it. The paint is not very thick, but it you can abrade the top layer of grunge off, you might get somewhere.
Or is this just an academic exercise? That board looks to be in pretty rotten shape overall.
Resistor color codes were just a rotten idea from the beginning, and are one of those maddening little archaisms that's still hanging around. Differences in human color perception can be very pronounced (even in person and not altered by a camera and computer screen). Diagnosable color blindness is fairly frequent, and pretty much everyone perceives color differently to greater or lesser degrees.
Stir in manufacturer variations in paint color, thickness, and application, not to mention the wildly varying effects of age, dirt, and heat, and you've got a ridiculous mess of mysteries.