Idk I already thoroughly explained one page ago why thaumas coat is overrated, with no counter argument other then the spreadsheet says so... However I explained as to how it can very much be overrated and why the spreadsheet can miss what a parser can see.
Actually, pchan's argument on this point is the easiest explanation: variance.
Manibozho body only has to deal with the basic variance in how much damage each hit does; Thaumas body has to also deal with the variance in how many extra attempts to swing are made. That extra variance means that there's a much larger range in how much damage you're doing at any given point in time, and individual observations (ie: parses) may show Thaumas as significantly better or worse than Manibozho under the same circumstances.
As a secondary effect on real-world usage, Thaumas also loses a small bit of its potential when fighting lots of regular mobs, since at least some of the time the extra hits are going to proc on the killshot round and be wasted. The spreadsheet only looks at the idealized form, though, and I don't have any estimates on how much is likely to be lost from that overage. A bit of napkin math indicates it may be as much as 0.5% of total DPS for mobs with 10k HP; double that for 5k HP mobs, halve it for 20k HP mobs, etc.
Leviathan.Dragonlord said:
»I don't know if the spreadsheets factor in the chance that additional hits may go over 100tp and therefore lose efficiency.
They do. Pchan made sure I wasn't lazy in that regard, and set up proper probability distributions.
The main local variance in that regard from the spreadsheet is the average over-TP value. I set it at 0.5 rounds, but that just comes from the best, most consistent values I've seen from my own (and a few others') parses for highly aggressive play. It's very easy for it to go higher, very difficult for it to go lower.
However it is, once again, an average value (and only an estimate, really, as well). A period of 0 over-TP compared with a period of 1 round of over-TP can greatly skew the apparent value in a parse observation. A quick look on the spreadsheet shows a 4% loss in DPS going from 0 over-TP to 1 over-TP.
However it is a simple fact that there is lots of things that cannot be understood by a simple formula. However you can account for some tp overage, the only real way to see this correctly is if you were to actually simulate rather than spreadsheet with dps and try and find how things will line up.
There are definitely lots of things that are abstracted away in the spreadsheets. *Most* of them can be safely ignored, I think, as long as you're using spreadsheet-to-spreadsheet comparisons. It may not match actual real-world data, but a configuration that's 5% ahead on the spreadsheet should also be about 5% ahead in actual usage.
However it also doesn't mean that you can't understand where those sorts of issues come into play if you have a decent understanding of the spreadsheet, and take the time to look at all the details it's telling you.
