It is one of those myths that persist, treating LEDs as light bulbs. It is probably maxing out the power supply too, which is saving the LEDs for a little while longer. A component that is normally rated for many hundreds of thousands of hours is now down to a thousand hours or so, if you're lucky. It's not doing the power supply much good either, its lifespan is being affected. The heat is telling you something.

Most folks design for fail safe. When it finally goes out all the LEDs will have been damaged, not just the one that burned out. At that point fixing the design will come too late for those components, including the power supply. The damage is cumulative.

You are speculating that LEDs have zero internal resistance which is also incorrect. A Thousand hours is less than 2 months, so I can know about that soon.

 

Then please attach the detailed datasheet.

I don't have such a datasheet here. These LEDs were NOS from the U.S.
Never had any use for them until now.

 

The minimum forward voltage of your 10 red LEDs is 1.8V BUT YOU POWERED THEM WITH 20V! Then of course their current was way too high without a current-limiting resistor.
What would happen if all 10 of your LEDs were the max of 2.5V and you fed them 20V? NO LIGHT!

Your meter is 1.2 ohms so it barely limits the current.

Chinese flashlight manufacturers buy millions of LEDs, measure the voltage of each one, label them and stack them so the current-limiting resistor can be small. You did not doo dat. Instead, your circuit is for ten 2.0V light bulbs.

18.9 volts.

 

And still no current measurement.........