Quickly made a map to debunk the myth that carpets are ‘6 times’ better than wood planks:
This doesn’t really visualize wood light setups (to verify how good wood light setups are you can use Flintless-simulator: Simulates flintless nether portal lighting), but rather simply visualize how much certain blocks catch fire from lava, spreads to other air surfaces, and burn out.
The truth is, the ‘ignite odds’ shown on the wiki aren’t exactly odds, but more rather ‘encouragement levels’. The true odds are computed much more (and I would say unnecessarily) complicated. The rate that all blocks that can catch fire from lava catch fire from lava however, is to my knowledge the same. Even setting the map to counting fire after 15 seconds, which is the average time of a wood light, I can only get that carpets in the 3x3 group counting ~1.8 times more fire than planks; In the 1x1 group where fire cannot spread across groups, that number drops to ~1.2.
In Wood Light ‘Standardization’ Solutions & Quantitative Measures to Decide How Good/Bad They Are, I swapped in carpets instead of planks and ran the simulation(note that this is not even possible because that requires floating carpets), and it showed that even in that case carpet light is only ~1.6s better on average and the time distribution have the same shape anyways.