-
One for the Dark Lord on his dark throne
In the Land of Mordor where the Shadows lie. -
There's this cryptic thread about the automaton by half. Seems he did make it for sil but it's extremely basic.Comment
-
Welp, this is what I get for not having looked at it for a few years. I'm way off base. There's a giant comment at the top of the file even that it's designed for Sil specifically...
However, yes, it is very simple. It doesn't deal with abilities, or item flags, or consumables, or running out of inventory space, or many more things. It is however capable of wandering around, finding a bow and arrows, and shooting enemies, which is somewhat cool. It
It might be possible to repurpose some bits of this to copy some of the functionality Brogue has around automatically going to the nearest down stairs and stopping if an enemy is sighted, which is rather nice as a UI feature.
More validly "unused" elements would be things like RBE_LOSE_MANA, RBE_UN_POWER, RBM_SPORE etc.Last edited by Quirk; January 14, 2022, 13:09.Comment
-
This is probably someone else LOTR's Beleriand's map project. Nick, what do you think of this map's accuracy?My best try at PosChengband 7.0.0's nightmare-mode on Angband.live:
https://www.youtube.com/watch?v=rwAR0WOphUA
If I'm offline I'm probably in the middle of maintaining Gentoo or something-Linux or other.
As of February 18th, 2022, my YouTube username is MidgardVirtuosoComment
-
It's not bad. It's based on CJRT's map for the published Silmarillion; I've been working from JRRT's "Second Silmarillion map" as it appears in Volume 11 of History of Middle Earth - here's an online version.One for the Dark Lord on his dark throne
In the Land of Mordor where the Shadows lie.Comment
-
I just dropped by and was delighted to see movement on Beleriand. Nick, you are entirely welcome to use any or all of Sil's ideas and/or implementation in Beleriand. It would be a wonderful home for many of these ideas.
As it forked off NPP 0.4.1, the code base pre-dates a lot of improvements that were made to V. And I'm sure I lowered the standard a bit more too, with some idiosyncratic ways of doing things (sorry Quirk!). One helpful thing is that I usually kept the rules simpler than is the norm in V. e.g. even behind the scenes calculations for monster generation etc was usually done in such a way that it would make sense to explain in a rule book. The complexity standard was something like AD&D 2nd edition, (whereas V allowed crazier stuff than even Gygax would have come up with, where you are dividing equations by random numbers between 1 and your dungeon level, and things like that).
Re the history of the code etc, I've just gone on a quest to see how much I've kept and … it's a lot. Heaps of old code and ideas files going back through all alpha versions and as far back as simple changes to edit files in 2001. I'll contact you (Nick) about how to get this historical archive to you.Comment
-
Nice to see you here
I was planning to email you when I was a little further along in the process. This thread has some useful ideas. One interesting thing is that Sil has the strict clock and sense of confinement, and I'm aiming to do almost the opposite.
I think you're being hard on yourself here. As you say, Angband code has improved a lot since NPP forked. Takkaria deserves most of the credit for that (both actual coding and the vision for what needed to be done), but also there have been a huge number of people who have pitched in and fixed particular bits of it.
I'm fairly deep in the Sil 'translation' at the moment, and pretty happy with how it's going. I am certainly seeing the virtue in your simplicity of rules, and it should translate into code that is in some places much cleaner and simpler than V's. As usual with this sort of task, it's pointing out ways in which the V code can be improved, and also some Sil features that could be imported (the hit display, for example).
This is excellent, I'll look forward to it. I already seem to be becoming some kind of archivist, so this should fit right in
One for the Dark Lord on his dark throne
In the Land of Mordor where the Shadows lie.Comment
-
I think it's inevitable that you won't produce your best code when extending a messy codebase - I am certainly not proud of the implementation of all my additions either, but the effort required to perform a clean re-implementation often feels prohibitive and so the existing pattern gets extended with all its warts, or hacked.
I do very much appreciate the consistency of the core rules, which I think has made the game much easier both for players to understand and developers to balance. I've certainly expanded my grasp of dice probability distributions through working on Sil-Q!Comment
-
Yes, I spent a lot of time thinking through some of those probability distributions and dice. The historical progression was something like this:
I didn't like the flat distribution on attack rolls in D&D (1d20+bonus vs AC), and I thought that 2d10 was better. I liked the idea of opposed rolls though, and at some point realised that 1d10+attack vs 1d10+AC was equivalent, but more elegant. It also generalised to some of the other skills I was thinking of and made me think that there could be an evasion skill that went up with experience (unlike D&D where, strangely, only attack does and AC tends to scale only with increasingly magical armour). During balancing, I realised that melee/archery/evasion were still too good compared to other skills and ended up giving them 1d20 rolls to balance them, adding some slight inelegance back.
I'd decided early on that I wanted to try a damage reduction system, as I noticed it allowed an interestingly multi-dimensional system for weapon choice vs different armour, and also for how these fit with different combinations of stats for the user. e.g. that you could make a system where if the opponent had light armour, you would be best with a light weapon against them (as it is sufficient to get through and has higher chance of hitting). This seemed interesting, natural, and modelling real world developments. It gets lost a little in the game as there are so many enemies and you can't micromanage it, but the system would also work well for a table top game where there were few, but high stakes duels against another person.
I started with flat numbers of damage reduction, as is standard, but then noticed that using dice for it (what I call 'protection') completes the duality between attack and defence, where you can have (+3, 1d10) vs [-2, 1d6] or whatever. And after the opposed rolls to hit, there are opposed rolls for damage.
As is often the case, it was a case of a few goals and quite a bit of time mulling them over and trying things out that eventually led to a very simple and elegant base system that really felt like a natural sweet spot in the space of combat systems.Comment
-
The opposed rolls really make the system. In particular I think the damage vs armour rolls are inspired. The old D&D rules are for me a bit of a failure both as simulation and on a tactical level (though in their defence combat is not too slow, and they inspired a lot of other people to do better), and Sil is a much more elegant and comprehensive attempt at modelling combat. I would class it above most tabletop combat systems, and in terms of tactical accessibility few other computer games have the same richness.
There are many elements that combine ingeniously. The malus for being surrounded and the morale and aggression systems are a couple more I'd highlight as significant influences on the gameplay.
That said, since we're on the Beleriand thread, there are a number of issues I have found over the years in balancing the game which would be useful to learn from if developing a new game.
You put your finger on a key theme with Melee/Evasion - which, to extend more broadly, is that while symmetry is seductive sometimes following symmetry leads to systems being pulled out of balance and the player experience can be pulled away from the sweet spot in order to make the character development screen look neater. I think there is sometimes a need to draw separating lines between how dissimilar systems work and stop forcing symmetry, and that your call with the opposed 1d20s was absolutely necessary.
In general, one size does not fit all with probability distributions. One issue with all normal distributions is that skill investment snowballs. Opposed 1d10s are particularly vulnerable here because it's quite a narrow distribution.
(If you have 6 skill points more than your opponent, you are already at 90% odds to win; one more takes you to 94%, one more 97%, one more 99%, then 100%. The first investment takes your opponent's victories from 1 in 10 to 3 in 50, the second halves this to 3 in 100, then cuts by a further two-thirds to 1 in 100, then your opponent winning is impossible. Increasing from 1 point ahead of your opponent to 2 conversely only takes your odds from 64% to 72%, i.e. taking your opponent from slightly better than winning 1 in 3 to slightly worse than winning 2 in 7.).
I suspect much of the time d10s are chunkier than would be ideal, because relatively small differences in skills can divide enemies into groups which are almost entirely neutralised by the check (for e.g. stealth, Lorien, Mastery) and groups which are quite dangerous.
Small differences in skills mattering a lot make it more challenging as the balancing developer to understand what range of skills players might be bringing to a particular encounter and hence what fair stats are for enemies, how good does a bonus on an item have to be before it's too good, etc.
Much of this is balanced by the variety of systems - e.g. the chunkiness of the stealth/perception check is lessened by the modelling of sound attenuation and other perception elements so the stealth experience overall is smoothed into a thing of beauty - but unfortunately where things most tend to break down is the late game, which would ideally be the tensest and most exciting part. This is very much the case with Evasion even though it uses d20s and much more investment is needed to more or less shut down enemies altogether.
"Sil is subtle", as they say. In other places also changing a single number can make for a world of difference. Balancing shortswords and daggers has been tricky for precisely this reason: the difference between a d5 dagger and a d6 dagger is substantial, and balancing so that a d5 dagger is potentially relevant to a fresh character and a d6 not overpowered to a stabber is a hard problem that has led me to drawing a lot of graphs.
In summary I would advise Nick to start with larger numbers and possibly wider distributions. This means a simple increment or decrement makes for a smaller percentage difference, and balancing becomes less fiddly. I would suggest also that symmetry often needs to be broken between systems that function very differently and it's maybe best to start by considering the needs of the systems and identifying where things are alike enough to overlap mechcanically rather than stretching the same mechanic over everything.
All this comes however from a place that could be considered perfectionist. Sil mechanics are honed to a degree few other games of my experience match, and if they could be honed further, it's probably only a few obsessives like myself who would notice or care.Comment
-
Yes, this is quite extreme. One of my aims was to reduce the numbers that appear in the game down to a manageable size. I think I ended up with numbers that are about half way between D&D numbers and Angband numbers (and much lower than, say, JRPG numbers). But unlike D&D, I wanted all weapons to be unique and not have any that are strictly better or worse than each other. And I needed to do this while having the numbers be thematic for that weapon (e.g. that they would end up being good when wielded by the kind of character that intuitively would be good at wielding them in the real world).
For some reason I seem to have been anchored to the Angband longsword at 2d5. So that is where I started. And through variation in the five numbers: attack, dice, sides, evasion, and weight, I managed to meet all my objectives. It wasn't easy, as these were also the numbers I needed to later tweak to adjust balance, leaving little room to move.
To top this off, I had various aesthetic ideas about how the numbers should look. I started with a triple of swords: the dagger (1d5), the longsword (2d5) and the greatsword (3d5), which seemed elegant. And then I decided to make all chopping weapons (e.g. axes) do even sided dice, while others do odd sided dice.
So there was indeed very little room to tweak things. It ended up a bit like a short poem, or similar, where it is highly constrained and every part has a purpose, and it becomes remarkable that it is possible at all. I love this kind of thing (especially the end product) as something to marvel over, but you are very much right that can be a lot of work!Comment
-
I wonder what odd taste for structural poetry is responsible for the 8d1 hammer in vanilla.Comment
-
This I consider one of the game's great successes, though as you say the tightness of the numbers can leave relatively little room to tweak things: most weapon types are viable right down to Morgoth (at least as artifacts, or smithed) assuming the character is built to take advantage of them. However, if I were embarking on a new project and wanted to make use of similar mechanics while preserving the goal of every weapon being unique, I would consider very carefully how many weapons I wanted the game to have before settling on the numbers.
Sil is definitely much like a poem. The constraints, of this type and others, have at times seemed torturous as I have laboured to balance a weapon or replace a skill. It's instilled in me a deep respect for the original vision that worked so fully with such a small and consistent palette.
I am curious to see what Nick will do drawing on Sil as an inspiration.Comment
Tweet
Comment