EPTI ran "suicide times" for every unit in both games that can attack itself (or rather, another unit of the same kind). I didn't even attempt to adjust for shield recharge/heal + repair/regenerate but hopefully the fact that I ignored it completely means it cancels out. I didn't even bother to work in stim, though I did calculate the differences in both games for tanks in their two modes and Vikings in SC2.
I ran two sets of calculations: one includes any buildings with attacks (including a marine-loaded bunker, not including turrets and spore obviously) and special/summoned units. The other includes only normal fighting units.
BW times were calculated using the 1/26th sec per frame/cooldown that I found a bunch of places. All stats are from liquipedia.
tl;dr: All the above says is that I have some numbers now that may or may not be useful.
Everything
Brood War
Protoss: The average kill time is 16.87 seconds, after 16.36 attacks.
Terran: The average kill time is 9.29 seconds, after 7.25 attacks.
Zerg: The average kill time is 9.38 seconds, after 8.73 attacks.
Protoss are tough bastards. Note that while their HP clearly is higher, attacks aren't much overall. Terran and Zerg really come out about equal – Terran may actually be more fragile. (Obviously medics beat hell out of this estimate.)
Starcraft 2
Protoss: The average kill time is 19.91 seconds, after 11.42 attacks.
Terran: The average kill time is 18.28 seconds, after 23.00 attacks.
Zerg: The average kill time is 14.23 seconds, after 11.33 attacks.
Terran are speedy bastards. Zerg are super-fragile.
Regular Units Only
Brood War
Protoss: The average kill time is 17.72 seconds, after 17.00 attacks.
Terran: The average kill time is 9.19 seconds, after 6.27 attacks.
Zerg: The average kill time is 11.09 seconds, after 9.88 attacks.
Protoss are still much tougher unit for unit. (Obviously this isn't taking into account the supply differences.) I can't tell if Terran just do a ton of damage, or are fragile, or both. Zerg average is thrown way the hell off by the Ultralisk since they don't have as many units as the other races.
Protoss: The average kill time is 19.90 seconds, after 11.00 attacks.
Terran: The average kill time is 13.11 seconds, after 21.83 attacks.
Zerg: The average kill time is 15.47 seconds, after 12.22 attacks.
The huge drop in kill time here for Terran is due to the fact that the Planetary and Auto-Turret stat out as super-tough against another one of themselves (and really in general). Zerg's goes up a little because the Broodling and Infested Terran are cut out.
In both games nothing really changes here. Obviously these straight averages don't really reflect common unit composition etc. but they do show a baseline that could suggest something. Not sure what though, other than to suggest that the initial post provoking this blog was dead wrong in saying SC2 units "die too fast" compared to Brood War. (Though stim is hellaciously better in SC2 which probably creates the impression so maybe it wasn't completely off base.)
One thing of note: BW unit cooldown times range from 8 (the Corsair) to 100 (Devourer), a spread of about 12x with an average around 31, 4x the lowest and about 1.5x the most common (22).
SC2 unit cooldown ranges from 0.225 (the Battlecruiser) to 3 (the Seige Tank in siege mode), a spread of about 13x with an average of about 1.45, right at or just under the most common cooldowns (1.4~ or 1.5).
Taking out these outliers, BW goes from 15 (Marine) to 64 (Valkyrie), about 4x. SC2 runs from 0.6 (Void Ray) to 2.21 (Mothership) slightly less than 4x. (Again, not accounting for stim.) SC2 features more specific variations (1.41 and 1.44 are both actual cooldowns) but feels like there's actually less overall variation. Though that could be just the fact that units which in BW saw wide use – corsair, reaver, sieged tanks – were noticeably different from the "standard" and I'm imagining stuff.
I really have no idea what this means, if anything. Any sort of detailed examination would have to balance for actual compositions, non-mirror rates, etc. etc. If I have to draw a conclusion on so little data, it's this: if SC2 is flawed (and I'm more and more thinking it's not, just young), it's not because of any radical overall differences from BW as far as the basic units go.
