The Feedback Issue – Which Weapon Should I Use?

You were to adventure in the world and not focus on min/max or optimizing or beating the numbers. You were there to group up to go forth and find adventure. Your mind needed to be on the world described, not on some die roll.

Me, attempting to sum up the early philosophy of TorilMUD

Potshot latched onto this quote in a past post where I was going on about changes in TorilMUD.

The context was around the fact that TorilMUD is exposing more numbers to players.

The game, based on the early philosophy I mentioned above, strove at times to hide what we might consider some of the most basic numbers in the game.  Rolling up a character required accepting stats that were not numbers, but just descriptions.  You might see “average” or “above average” or “mighty,” each of which mapped to a range of values.  The numbers were eventually revealed once you hit level 20, by which point you were presumably committed to a character, though if this was you 4th or 5th character, you probably had enough equipment stored away to twink them out, and enough knowledge of where to go, to get them to level 20 in maybe just 8 hours of game time.

That started to change over the years, especially during the latest incarnation of the game.  And the changes were primarily justified as being about providing feedback to the players.

The first thing to change was how you could check on your level progression.

Back in the old days, you had to go all the way back to your class guild master and check on your experience, which resulted in messages like this:

The great druid Kaladan is ready to show you how to become one with nature.
Your guildmaster says ‘You are still a very long way from your next level.’

I think that meant I was between 20% and 29% into my current level.  There was a different message for each 10% graduation per level.  And while some of the messages were more obvious than that… before and after half way said just that and for the last 10% your guild master grinned in anticipation… it was still a pain to travel all the way back to town just to get a reading on your progress.

So that changed to a text version of a progress bar, then to a simple percentage read out, and, just recently, slaying mobs began reading out both an experience point value and a percentage like this:

You beautifully slash a burly sailor into two parts – both dead.
a burly sailor is dead! R.I.P.
A burly sailor slumps to the ground.
You receive 40,573 XP (1.07%) experience.
Your blood freezes as you hear the rattling death cry of a burly sailor.

So there you go.  It is now possible, 19 years after the MUD first started, to directly assess the value of a given mob.  And the “exp” command tells you how much you need to get to the next level.

You are 5,101,956 XP (94.48%) away from your next level.

The problem is that we have now moved from levels being something of a mystery to levels becoming a mathematically precise certainty, which is a clear step away from the original philosophy of the game.  The next step would seem to be to expose hit point values and damage rolls.  Right now those are still hidden with verbiage.

You parry a burly sailor’s lunge at you.
A burly sailor’s attack only grazes you as you maneuver your mount!
A burly sailor slightly wounds you with his average hit.
Your mighty slash slightly wounds a burly sailor.
Your attack only grazes a burly sailor as he dodges aside!
Your strong slash barely wounds a burly sailor.
< 400h/427H 210v/210V >
< T: Kigev TC: few scratches E: burly EC: small wounds >

But is that the right direction?  Must we always move towards exposing more numbers?

Certainly that is the easiest way to express feedback in a system that is made up of numbers.  And if you are going to try and hide numbers, you have to come up with an effective way to provide feedback on some things that we might otherwise not consider, such as how to tell which weapon you ought to be swinging.

Weapon comparisons have been done with numbers… which pretty quickly got summed up in DPS ratings… for a long time now.

Have shovel, want mallet!

WoW weapon comparo

But could you do it without numbers.  Could you look at a weapon, equip it, maybe try it in a fight or two, and get enough feedback to say whether or not this is what you out to be swinging.

I decided to check TorilMUD to see if perhaps weapons gave enough description for that sort of thing.

Certainly some do.  The description for my Paladin’s holy avenger lets you know that this sword is something special.

This heavy sword has been crafted out of an unknown metallic alloy, the exact nature of which is known only to the gods. The long blade gives off a soft and warming radiance, even as the edges glint dangerously. A hilt long enough for two hands to grip firmly has been decorated with kingfishers and the pommel is crafted to look as though a dragon maw is gripping a brilliant pearl. Flaring crossguards sweep up, masterfully tapering into the appearance of talons that meet the bright blade.

After the long quest to obtain it, you were probably pretty sure it was going to be hot stuff in any case. But what about further down the food chain?  I decided to look at weapons that new players might pick up, to see if I could correctly pick the best weapon by looking at the description.  In order to limit the range and to keep to places I knew well, I focused on the areas outside of Leuthilspar, the elven starting area.

In some of the old haunts I was able to pick up five weapons from various mobs to see what their descriptions said.

Name: a bronze sword
Description: The sword is fairly small yet broad, with a thick leather handle. It looks perfect for close in encounters.

Name: a small sword
Description: The small sword seems to have an inscription of some sort.

Name: a long sword
Description: you see nothing special

Name: a cudgel made of stonewood
Description: This blunt, short club is made from a special type of wood which is hard as stone. Crafted by the special skill of the faeries the club is impervious to damage.

Name: a wooden spear
Description: This wooden stick is almost but not completely straight, it is about two feet in length. Sharpened to a point it makes a crude but usable weapon as demonstrated by the dried blood on its tip.

So, given those five choices, which would you choose, assuming you have chosen the warrior’s path and are thus likely not to face any class restrictions?

Actual stats after the cut.

So, which one do you pick up, young adventurer, given that you are new and poor and cannot afford a scroll of identify to help you out?

The one with no description at all, the long sword, seems to be the best choice.

It is the lightest weapon, which counts for something in a game where weight imposes penalties, and is tied for best die roll.  The bronze sword has the same die roll, and is only a little bit heavier, so is pretty close.  I feel better having always gone for the bronze sword back in the day.

Here are the stats in descending order:

Name ‘a long sword’
Item type: WEAPON
Item can be worn on: WIELD
Item will give you following abilities: NOBITS
Weight: 8
Value: 600
Damage Dice are ‘1D8’

Name ‘a bronze sword’
Item type: WEAPON
Item can be worn on: WIELD
Item will give you following abilities: NOBITS
Weight: 14
Value: 650
Damage Dice are ‘1D8’

Name ‘a cudgel made of stonewood’
Item type: WEAPON
Item can be worn on: WIELD
Item will give you following abilities: NOBITS
Weight: 14
Value: 20000
Damage Dice are ‘1D6’

Name ‘a wooden spear’
Item type: WEAPON
Item can be worn on: WIELD
Item will give you following abilities: NOBITS
Weight: 2
Value: 3360
Type: Spear Class:
Simple Damage: 1D5 Crit Range: 3% Crit Bonus: 3x

Name ‘a small sword’
Item type: WEAPON
Item can be worn on: WIELD
Item will give you following abilities: NOBITS
Weight: 3
Value: 60
Damage Dice are ‘1D4’

The wooden spear is kind of an odd bird.  If you were a rogue and could backstab, that would be the better choice.  Likewise, if you were a cleric, the cudgel would have been your only option.  But for the warrior, neither would top the list.

So, sadly, despite the original intent of the game, you really needed the underlying numbers to make an informed choice about your weapon.

But does it have to be that way?

What would it take, what sort of descriptions would be required, what sort of feedback would be needed to avoid numerical values?

In real life you would pick up these weapons and would have a sense about which one would suite you.  How do you translate that sort of feedback into a game?  Would we have to de-emphasize equipment progression completely to do this sort of thing?

16 thoughts on “The Feedback Issue – Which Weapon Should I Use?

  1. Knug

    What would be the point of the detailed description? Someone is just going to post a spreadsheet detailing the description vs stats online anyway.

    Some games do add descriptors for weapons based on the various bonuses to that weapon. I distinctly remember Dungeon Siege, where “Jared” meant it had electricity or something, with an adjective giving relative strength.

    However, someone did have a “translator” up pretty quickly after the fact.

    In EVE, Wormhole types introduced with the Apocrypha expansion were identified, with rather vague values – it was supposed to be mysterious, risky, the great unknown. It didn’t take long for the Internet to be sprinkled with programs and tables outlining all the details for every unique wh identifier. Exploration of the unknown looses its value, when every possible combination has already been found/documented/detailed/analyzed to death.

    In this day and age of instantaneous knowledge (or at least instantaneous facts) depending on the unknown, the vague, or the approximate as gameplay additions aren’t enough. The spoilers are out there.


  2. Wilhelm Arcturus Post author

    @Knug – “What would be the point of the detailed description? Someone is just going to post a spreadsheet detailing the description vs stats online anyway.”

    The presumption, which I did not state clearly, is not to go back and do something like “retrofit WoW” to add descriptions. It would be more along the lines of some new game in the future dispensing with such stats.

    Can you make a fantasy MMORPG without resorting the the massive numbers games?


  3. bhagpuss

    I certainly believe you could design an MMO that only gave descriptions for everything and never numbers. In fact, unlike a text MUD, you wouldn’t even need written descriptions for things like weapons – assessment could depend entirely on appearance. Personally, I’d jump at the chance to play that game.

    The problem would be secrecy. Many players would see it as a challenge to find out the numbers and you could hardly find a worse audience to try and keep this kind of secret from than millions of tech-savvy Gamers, many of whom actually do this sort of thing for a living. If you were able to design things so that the code couldn’t be cracked from outside you’d still have the problem of leaks from within and games development has something of a history with that.

    We do have examples of companies that do it, or try to. Square Enix are notoriously cagy about the exact numbers behind the Final Fantasy MMOs, for example. There was that whole thing about weather/times of day affecting crafting which went on for years. I think the people behind Wurm Online do it too. SoE used to do it with Everquest and I liked it much better when they did, but market and player pressure wore them down over time.

    One thing, though; if a developer did try this they’d need to be a lot clearer with the descriptions than the TorilMUD ones above. I’m guessing they look a lot more intuitive to you because you played the game for years but to me as a complete novice I would have ranked them like this, and for these reasons. Best at the top, worst at the bottom:

    A small sword. It’s got an inscription! It *must* be enchanted, otherwise why mention it? And because it’s such an oblique, subtle hint it’s got to be a really good enchantment!

    A cudgel made of stonewood: it was made by faeries for heaven’s sake! And they used a special wood! And it can’t be damaged! How cool is that? I’m going to be using this baby for levels and levels!

    A bronze sword: okay, it may not be magic but it is “perfect” for close fighting. Perfect. That’s the best you can get.

    A long sword: It’s boring. It’s so boring it actually says so. “Nothing special”. Now given it does say “You see nothing special” and that could be a hint. Maybe if you Identified it you would see something special. But maybe what you see is what you get. Hmm. I wouldn’t sell it until I’d identified it, that’s for sure.

    A wooden spear: it’s a frickin’ stick!


  4. Wilhelm Arcturus Post author

    @Bhagpuss – Ideally, I wouldn’t want to just hide numbers and leave people in ignorance, but give people an informed method of choosing gear that did not directly reveal the underlying math. It would be great if you picked up a sword and felt that it was a good fit for you, or awkward given your strength, class, race, or whatever. And maybe an innate compare ability when you have two weapons so you’ll get a feeling for which is better.

    EverQuest always had some numbers. There was the damage rating and the speed of the weapon listed back at launch, so you had to do some minor math to get a comparable stat.

    You interpretations of the descriptions are almost exactly what I would have said, aside from the wooden spear. And the spear was different because I knew from the past that it was a deadly piece of wood that once had a much high die rating. It was wielded by a short swarthy faerie that would flee to another room, hide, and then backstab you to death if you were not very careful. But back in the day I latched right onto that “inscription” on the small sword. I knew it just had to mean something. It didn’t.

    Meanwhile, “You see nothing special” is the default that comes up when the person who created the item failed to enter a description.

    The problem with the descriptions is that they were all done individually without reference to anything else in the game. There is no normalization in that regard, although the stats have been normalized over time.


  5. Celendus

    A “compare items” feature might be useful; place or link both items into it and it would say, for example, “Longsword is slightly lighter than Bronze sword (<10%)" "Longsword is more than twice as deadly as short sword (250% to 300% base damage)". I'm sure someone out there would compile a database of comparisons and create a whole system using a cheap flint dagger as the base unit, though.


  6. Jonathan

    And… honestly, isn’t that more realistic? I mean, as much as we want things to be mysterious and for us to discover them… do you buy new hardware you want to depend on sight unseen, or do you do a bit of research and read some reviews?

    To me, the real problem isn’t that we can’t get to the “right” answer quickly. Its that there is a “clearly correct” answer to begin with. In reality, a sword that is best for one person isn’t the best for another one… arm length, strength, training, fighting style all play a role. Here, well, 1d8 is just flat better than 1d6. It just is.

    The real world is mysterious because it is complex. I’m not really sure how we model that.


  7. Knug


    If we are considering a brand new game, and have stated the challenge that we don’t want numbers brought out, what kind of game is it ?

    Mathematics is the basis of all simulation, and what kind of adventurous game are we building if it is not a simulation to some degree? If we are to have immersion, our world must be consistent within its own ruleset. How do we define that rule set without a matching mathematica/logicial construction?

    If our game’s world is underpinned with mathematics, there will be a strong desire to unearth the numbers. Our human nature of competition and curiosity will drive the need to derive those formulae, if nothing else, to better our position with others.


  8. Whorhay

    What Jonathan said, so long as there is actually a clearly best weapon over the others given that you can use them all, and that the game is all based on numbers.

    EQ wasn’t quite as straight up about the numbers as you might think. Some classes actually got bonus damage per hit, Rangers specifically, which you never got to actually see a numerical value for. Also there were damage caps per hit based on character level. So given two weapons with relatively close damage/speed ratios you’d be best off going with the faster weapon. There were of course exceptions because you could use strategies like a Paladin “jousting” with a big slow two handed weapon.

    So long as our games are relying on numbers systems for damage and whatnot then the numbers are what I want to see. If the game is much less reliant on a bunch of math but instead strategies and technique matter then I’ll be happy to go with weapons without listed stats. Incidentally that ‘wooden spear’ sounds more like a large tent stake, seriously two feet long, you might as well use a willow switch.


  9. Knug

    If the purpose of our game is to have an automated game master (the current term for a good old dungeon master) that lets thousands of players take part in an RPG, then we will need to have some system for it to do it. We can cover up the numbers as much as we like with things like “feel” “intuition” “comfort” etc. But those that play the game will associate things like “sword feels comfortable in the hand” and “wow, those kobolds died really easy this time, what changed?” It won’t take long before players set stats on their own (even if made up ones, that say use a starting knife as ‘one’) to create their own way of simulating the game world we’ve given them. After a time, their analysis will have created a parallel statistical system that will approach the ‘true’ stats behind the world, mirroring the way our mathematical model of the natural world becomes closer and closer to completely explaining all in our world around us. They won’t need the “actual” programming/data set, they will create their own to predict which weapon to use !

    Oh I should also add that when I was DM’ing (back when D&D was the only MRPG game about) the rules implied that the DM did all the rolling for NPC’s out of site from the players. The point was that the story being woven with all the players including the DM was more important than the numbers being rolled. A good DM was able to simulate the entire world in such a way as to keep it consistent. Random bad die roll indicates an NPC should die at the wrong time? Ignore it ! If nothing else, it screwed up the way players modeled the world, trying to find out the unique rules I built into it. By keeping that random element (which never negatively effected the ultimate outcome) it prevented the game from becoming too predictable.


  10. Vatec

    On a related note, Asheron’s Call launched with a complex magic system that required experimentation in order to “research” spells, with the complexity increasing with the level of the spell. There was also a “spell economy” that was intended to encourage keeping ingredients and formulae secret (the more often a certain spell was cast server-wide, the less powerful its effects).

    The basic ingredients for all major spells were published online by enterprising players not too long after launch. You still had to “research” the taper component for higher-level spells, but at least you knew which scarab, plant, oil, etc. you needed.

    Then some bright person figured out the pattern to the tapers. And that person then wrote an application that helped you research your characters’ individual taper patterns as efficiently as possible. Some players bemoaned the “dumbing down” of the game, but a huge number of players downloaded the program and started playing mages. “Balancing through tedium” simply didn’t work in this case.

    The problem with any such system is that a few smart, inquisitive players with the “information wants to be free” mentality will break it as swiftly and thoroughly as they can. Quite frankly, it’s simpler less work to provide the numbers and let the players run with the information.

    PS: Never did understand why Funcom games tend to have damage ranges with bizarre fractional elements, like 42.05 to 87.32 or 150.01 to 185.03, though. Wouldn’t it make more sense (and provide a cleaner UI) to use integers? Or at least to display integers?


  11. Matt

    It’s not unheard of in games to dislike the ‘feel’ of a weapon, such that one prefers a numerically inferior one. The way this manifested most often in WoW was preferring certain speed weapons. Of course the hardcore min-maxers went for the numbers, but the mass of normal players would make a more ‘feely’ decision sometimes. The real world is full of these sorts of complicating factors that don’t necessarily translate to numbers, but games aren’t quite there yet. In fact, WoW just made all weapons within type the same speed, removing this level of complexity.

    So I would say that what you might want is not a hidden math simulation, but a ridiculously complex one involving a huge number of factors that no one could possibly follow without a 50 page guide, meaning most people just shortcut with their intuition as they do in real life. Now whether this would work or be at all worth the trouble I can’t say.


  12. Potshot

    The key as I see it lies in the relationship between the player and his environment. If there is sufficient variability in characters, classes, and for wan of a better term, personal progression mechanics, all vaguely described, then a game could provide feedback to the player after determining the player’s proficiency with that weapon given his or her class, race, level, skills, stats, etc. the goal being that at any given time for any given set of circumstances, the “correct” answer is almost unknowable and sufficiently unique as to be nearly irreducible to the One True and Correct answer.

    Buried beneath of course are the mechanics, but as the modeling would be dynamic one could grow into and out of a particular weapon or ultimately come to the conclusion that your character will never be able to effectively use such and such.

    Spoiler sites would hopefully become little more than guides of lore rather than simply illuminating the shining path.

    A man can dream can’t he?


  13. bhagpuss

    One solution to the “problem” of players figuring out the machanics (cf the excellent Asheron’s Call anecdote above) that I’ve often mused on would be to design the game so that many of the underpinning calculations and constants could be altered on the fly while the game was running. Then, rather than allocate this task to some kind of autonomic algorithm, actual human beings are employed to make changes as and when they see fit.

    In other words, each server and/or the entire gameworld has one or more GMs who fiddle about with the underlying physics. They wouldn’t need (or be able) to operate like an actual GM, taking action according to what players are doing at any given time; they’d simply change the parameters within those players could operate.

    For recipes that need to be found out by trial and error, for example, rather than have them all fit some predefined formula a dev could come into work on a Tuesday having had an idea on the commute in, slap a dozen completely new recipes into the database, take out half a dozen others and leave players to find out what had changed. Then by the time someone chanced on the new combinations and word began to get out another Dev could have tweaked things a little and all that received information would be slightly, or completely, inaccurate.

    It would clearly require some checks and balances and good management practices, and it would need to be done in a new MMO where players understood the environment they’d be operating in, but I think it would not only be feasible but probably not particularly expensive to operate if the tools were constructed well enough.


  14. *vlad*

    Back in the early 80s when I played AD&D, my Main used a Bec-de-Corbin, a big clumsy mace. Now it certainly didn’t do very good damage (2D4? I really don’t remember), but hey, RP over stats any day! Who needs a +5 Longsword when you can bash people with a big lump ot metal?

    I don’t like open stats; all that leads to is dissatisfaction with your current weapon/armour etc, and the need to ‘gear up’ becomes your focus rather than the story.


  15. mbp

    I am not in favour of hiding information that players need to enjoy the game but I do understand how immersion breaking it it when everything is reduced to a table of numbers.

    Perhaps this dilemma could best be solved by using more sophisticated weapon models. When a swordsman picks up a blade in real life he doesn’t calculate how much blood per second it will let. He will however pretty immediately get a feel for the length of the blade, the weight, the balance, the sharpness, the quality of the steel and so on.

    Imagine if each of these parameters contributed in a different way to the behaviour of the weapon in combat. Instead of reducing everything to a linear min-maxable DPS scale you would have an large array of possibilities. Some of the parameters (like length and weight) should naturally be given numerically bit others such as sharpness and balance might be grouped into a small number of categories. .


    Dobbin’s sword of cleaving is 1 metre long with a rounded point and a very sharp edge It weighs 6 kg and is well balanced.

    Noddy’s rapier of .pricking is 1.1m long with a very sharp point and a blunt edge. It weighs 3kg and is very well balanced.

    Those descriptions are entirely believable within the context of the game and yet they give you useful information that helps choose which weapon to use for a given job.


  16. Knug

    @vlad @mbp If ‘open stats’ aren’t given then players will develop and post their own stats, leading to the ‘need to gear up’ and or the classic max/min behavior.

    The greatest challenge for RPG is the RP part. With the advent of the computer as a GM, which can handle all the numbers/charts/odds/statistics you can throw at it far easier than any human, the development of the genre lead itself away from role playing and into stats based gaming.

    For many gamers out there, RPG means game where you have player stats. The concept of playing a role within a game environment is mostly lost. Very few folks actively RP the RPG they play. In the case of MMORPG, the fraction of RP’ers is quite small. Amusingly, one of the strongest reasons is peer pressure from other gamers, who want the action and stats of an MMO without the RP part.

    So long as humans are both curious and competitive, an RPG game will have stats assigned to it, either openly by the developer, or publicly by the players.


Comments are closed.