Conquest of Elysium 3

Conquest of Elysium 3

Marlin 1 May, 2016 @ 3:50pm
Extracted from coe3.exe: Class and Recruitment data
Update February 2018: The download link, below, is now to new data fresh from CoE 3.26!

Like before, the format is c3m files (plain text files) with pseudo-mod code. And, for the best reading experience, I recommend a text editor with CoE syntax highlighting, such as the schemes I have made for Gedit and Notepad++ – both of them incidentally now updated for v3.26 – or, if nothing else, the syntax highlighting of just a Unix shell script. (The last one will at least make comments and string literals stand out from the rest.)

Whatever you do, I would advise you don't use the Notepad that comes with Windows. (If you're on Windows, you'll be much better off with Notepad++[notepad-plus-plus.org].) Microsoft Notepad and CoE modding (or just about any serious text processing) are two things that simply don't belong in the same sentence.


The rest of this post is what I originally wrote in May 2016:
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
So...

I finally got around to take a good look at what the class and recruitment data in the executable actually look like, and found they weren’t at all as horribly indecipherable as I had feared. In particular, data seem to be generally stored in fixed-size blocks, thankfully also often begun with some telltale plain text string. Which greatly facilitates comparison between blocks, in order to figure out what each byte and bit might represent. Numbers seem to be mostly 4-byte (signed) integers, each beginning at an offset evenly divisible by 4. (When using a hex editor, I find it helpful to have it display bytes in groups of 4.) Although there are a few 2-byte integers too.

The ideal, in my view, would have been human-readable text files (e.g. like mod files in format) with the definitions of all the stuff in the game, classes, units etc. In their absence, though, I believe things could have been worse than they are.
________________________________________

I was eventually even able to put together some Python scripts to extract the data into... Not meant to be used as actual mods, of course. (There aren’t mod commands for all the data in the executable, and I don’t understand all of it anyway; see the comments.)

I could now kick myself for not doing this earlier. I did in the past actually waste some considerable time trying to establish only some of this through arduous tests.

For the best reading experience, I would recommend a text editor with CoE syntax highlighting, such as the schemes I have made for Gedit and for Notepad++, or if nothing else the syntax highlighting of just a Unix shell script (will at least make comments and string literals stand out from the rest).
________________________________________

Notes:
  • The file on standard recruitments (standard troops, standard wizards, paid heroes and assassins) is not extracted from the executable, as I cannot there identify the data for it. It contains guesstimates based mostly just on tests that I also now got around to perform. I now feel fairly confident this is how it really works, though there is almost no way to be entirely sure.

    (Again, I wish I had done this earlier. There have been times playing this game when the capturing of extra libraries felt almost like a pointless ritual based purely on superstition, as I had so little an idea what the actual probability bonuses were.)

  • The other files contain data actually found in the executable – complete with any errors there might be. Not that I expect there to be many errors, of course, but user Ridiculus recently discovered what seems to be a corruption of recruitment data for the witch (which must have happened in a patch).
________________________________________

For the curious, interested in a comparison with CoE4, I posted the corresponding CoE4 data in the CoE4 forum.
Last edited by Marlin; 14 Feb, 2018 @ 6:26am
< >
Showing 1-2 of 2 comments
Marlin 4 Feb, 2018 @ 2:36pm 
The effect of temples and libraries on recruitment chances

The effect of temples and libraries on recruitment chances was always very vaguely documented. (And there has been times when I wondered if they had any effect at all.) In fact, it may have been the deepest mystery of the game to me, ever since I started playing CoE3. So now that I finally think I have some answers, I felt I had to present them in their own post. (I have previously posted this also in the CoE4 forum, but I think it should be here, in the CoE3 forum, as well.)
________________________________________

Having extracted the class and recruitment data from the executable, I knew the base chances of all recruitment entries, as well as exactly which ones of them were boosted by temples or libraries (had the templerec or libraryrec properties/flags). It also let me know what the recwizchance class property was for every class (=0 for most of them), and through some fairly extensive testing I was able to finally nail (I think) exactly how that property works, i.e. what the base chances of standard wizard offers are.

Still, the exact effect of temples and libraries on recruitment remained a bit vague.
________________________________________

Now, finally, I decided to go through with the drudgery of testing how the temples and libraries of CoE actually work.

The results:
  • In CoE3:
    1. It seems that each temple and each library can give up to a +3% boost to the chance of a recruitment entry with templerec or libraryrec, respectively.

    2. However: If the base chance of the entry is 2% or lower, each temple/library only adds about +1% to it. If it's 3%–8%, each temple/library adds about +2%.

    3. No more than 5 temples and 5 libraries count for a recruitment bonus. If you have more, all of them will be shown in the recruitment window, but only 5 of each will count.

    4. The boost to the chances of standard (human) wizard offers seems to be the step size of recwizchance, 2 per mille, per library – again capped at 5 libraries. Which would give a maximum of +1% per standard wizard from libraries, or about +6% total to offers of any standard wizard type.

  • In CoE4:
    1. Each temple appears to give +4% to the chance of a recruitment entry with templerec.

    2. Each library point appears to give +3% to the chance of a recruitment entry with libraryrec. (Meaning a standard magic library now gives a +6% boost, and the home of an Archmage, what’s in CoE4 is known as an Academy of Higher Magic, gives 9%.)

    3. The boosts from temples or libraries are both capped at +20%. So no more than 5 temples will have any effect (like in CoE3). While up to seven library points count for boosting entries with libraryrec (the last one a bit less than the first six).

    4. The CoE3 reduction of temple/library boosts, where the base chance of the recruitment entry is low, does not seem to be there in CoE4. Establishing this for sure would take more tests than I care to make, but I haven't at least seen any reductions in CoE4 as sharp as in CoE3 – not even at a base chance of zero.

      In other words, while the effects of temples and libraries have been increased overall in CoE4, compared to CoE3, this is particularly true where recruitment chances start low.

    5. Standard (human) wizard offers finally: The effect of libraries on these is a bit confusing in CoE4 (less straightforward than in CoE3).

      • It seems that the boost per library point to the chances of wizard offers is slightly higher than the step size of recwizchance (which in CoE4 for most wizard types is 1 per mille). Possibly 1.1 recwizchance steps per library point?

      • Up to 10 library points seem to count for this purpose. Which is more generous than the cap on boosts to offers with libraryrec in CoE4, and also (normally) more generous than the cap on library effects in CoE3.

      The recwizchance step size is for most standard wizard types only half of what it is in CoE3, of course. But the maximum boost per wizard type is still comparable, if not bigger, and the number of standard wizard types is almost tripled in CoE4. So it seems you can in CoE4 get as much as one offer of some standard wizard about every fifth turn, on average, if you manage to capture 10 library points.
(Increases to recruitment chances are given in percentage points added, of course. Or, put another way, the individual chance of each temple/library being a source of the offer in question is 1, 2 or 3 percent. There is no multiplication, no percentage of percentage involved.)
UncleYar 4 Sep, 2021 @ 10:56pm 
Thanks for this and the data in the other threads. I discovered I like the combat in CoE3 more than the 4th and 5th installments because of the lack of friendly fire, so now I'm looking for mods for this game, and perhaps doing some light modding myself.
May I suggest uploading to Github the python scripts you used to extract the data?
< >
Showing 1-2 of 2 comments
Per page: 1530 50