the consensus is that destruction magic does insufficient damage. the specifics have been well described in various other threads, however, no fix has been suggested.

short version: use poly not a const for damage
long version:

weapons can scale with the character as they increase proficiency in smithing. perks further the advantage of using physical weapons. so magic damage should consequently also use a scaling damage based on the destruction skill. i believe it would be advantageous to have magic damage affect a percentage of enemy's health rather than an absolute value. percent enemy health damage will be abbreviated pehd.

pehd = ( [destruction skill] / 100 ) ^ n

because destruction will max out at 100, destruction magic will do 100% damage at 100 destruction, regardless of n. essentially, all destruction values are normalized (thus the /100). the exponent ensures that at destruction = 100, the return value will necessarily always be 1. this allows freedom to tweak without worrying about specific parameters does not have the scalability problems that linear functions tend to have - especially if difficulty is modulated. in order for this function to never reach 1, you can also set [destruction skill] -= 1.

for 0
for n>1, you will start at no percentage, and only at the very highest levels of destruction will you see rapid gains. for example, at n=9, there will be no observable percentage gains until destruction is 77 - at which point it will only do 10% enemy health damage. this is most viable if a spell retains its damage and does an additional pehd. usually around expert level destruction is when players expect to be gaining more out of destruction, at 92 pehd=50%. this scale can be easily changed by simply modifying n.

the domain is all real numbers - at negative values, pehd will result in outrageously high numbers. however, if the game somehow calculates a disease or exhibits wonky behavior, this function will not throw an error (for example, casting on a dead enemy with 0 hp).

this does not gaurantee a 1 shot kill. enemy resistance should affect the total damage done. halving the pehd to max at 50% would also work - a 2 shot kill if enemy is not resistant. this would of course, also apply to followers as well, encouraging the use of magic resistance gear.

furthermore, you can set n to the difficulty. to formally define difficulty, master difficulty is d=2, enemies do d times damage to you and you do 1/d damage to them. at low difficulty, you'll see that the percentages start higher and scale faster, but ultimately max out the same as for master. on extreme difficulty (for example, if d=100) you will see no gains whatsoever until destruction is 99 (pehd is only 40%) and only at 100 will you see pehd=100%. this means more training and more conservative play on master - what is expected on higher difficulty, yet the game mechanics are still retained.

examples: suppose there's a draugr deathlord with 1400 health and 50% frost resistance. at destruction 75, thunderbolt does 90 + 56% = 90+784 = 874. the next thunderbolt kills him. 2 shots against a non-resistant enemy. by comparison, ice spike would do 25+56% = 0.5*(25+784) = 404.5. it now takes 4 shots against a 50% resistance character (as expected). weaker creatures will still be 1 shot killed, since anything with less than 160 hit points will die instantly, which is expected; any time the raw spell damage is greater than the remaining percent. it won't be until destruction 99 on d=2 to 1 shot kill the same deathlord with thunderbolt. it will always take at least 2 shots with the incorrect element.

furthermore, this would allow spell behavior to increase spell use. ice spike retains its usefulness compared to ice storm in a group of friendlies mixed with enemies. ice spike at 25 damage, however, would not instantly kill a deer. the 25 damage would not make up the remainder, the characteristic of ice spike as an elementary spell would still be apparent. against immunities such as dwarven constructs or dragons, the incorrect element will do no damage, as expected. furthermore, spell damage only increases during battle - smithing can be arbitrarily increased in absolute safety (as well as poisons). this means its increased effectiveness in consistently more dangerous situations based on player skill could be rewarded. linear scaling by multiplying the damage makes all d<1 too easy and does not have an observable impact on d=6 (oblivion difficulty).