Disease Chances in AD&D/OD&D

[oakesspalding]

Okay, so I have this feeling that many of you are going to ignore or be hostile to this question because you "don't use that rule" or prefer to "just wing it" or whatever . But I'm interested in your thoughts on the chances for randomly becoming infected with a disease.

OD&D's Blackmoor has a fairly nifty section on disease, including a chart that gives you a "% to catch" stat for each disease. However, that stat is based on "whether the disease is present," which sort of begs the question.

AD&D has an inferior breakdown of diseases - suddenly they're all unnamed and annoyingly abstract. However, it does have a fairly robust mechanic for determining the random chances for actually getting a disease. The base chance is 5% a month, going down to 2% or 3% in, say, cold climes, but increasing to 10% or more in warm climes, mosquito infested swamps, crowded cities and the like. In turn, a random disease will be fatal approximately 12% of the time.

That means each character has a roughly 45% to 70% chance each year of randomly getting a disease, or a 5% to 10% chance per year of perishing from a disease. Obviously, this assumes you don't have access to any Cure Disease spells or similar.

I've designed a mechanic that approximates this, though you roll per day, rather than per month (the monthly and yearly rates work out to about the same).

The question is, are the above %s right?

I have players rolling 2d6 per DAY, with a 12 representing the chance of getting a disease, though often that "disease" will be "no effect." (The average chances are still 45% - 70% per year with a 5% to 10% mortality rate per year).

So should they instead try to get an 01 on percentile dice (this lowers things to 15% to 25% per year, or a mortality rate of 2% to 3% per year)?

Or should they be rolling 3d6 for an 18 (7% to 12% chance per year of disease, or a 1% to 2% mortality rate per year)?

How long is a campaign? How long does a character take to reach mid-level or high enough to have ready access to Cure Disease spells? What's the proper chance for being infected by random diseases? If it's too high, then characters will die too early from silly random diseases. If it's too low, then you're taking the fun - in terms of the "fun" of diseases - out of it. What's the point in having them if they so rarely happen?

Thoughts?

[howandwhy99]

I think in terms of thresholds. That 5% is significant enough for science and a d20, so it's good enough for my default threshhold. I don't believe there should be an annual 5% chance of your character dying rolled for each year prior to Starting Age. I think diseases should typically be less awful and less likely except in very squalorous conditions.

I'm down with a behind-the-screen roll for lengthy stays in clean areas, weekly or daily rolls in icky dungeons, and spot checks for environments where diseases thrive: like the mouths of rats for instance.

I don't tell players the results, but they can discern later as symptoms arise and penalties are taken. Let them deal with it or not, but discerning diseases and curing (most) without divine assistance should be possible IMO.

The only reason I'd keep the lengthy rolls is to account for such deaths in the large populations, which will have a small number die due to "normal disease", IOW not a plague event. For D&D players that roll would only happen after extended non-adventuring. I mean, PCs routinely have "sewers dungeons" and other nasty filth to deal with, so their game really is about specific location-based checks, like poison.

[hamurai]

I've never had my players roll for diseases just because they woke up or lived through a year. I've only ever checked if they caught something when living conditions deteriorate, meaning during a siege, plague, famine or some natural disaster like a flood, or when they've been down in a dungeon for a lengthy time without any chance to wash or when they had to eat rotten food and rest in really filthy environment. In these cases we did often use the d20: roll a 20 and you caught a disease, but in that case it wasn't just a cold but something more dangerous according to what caused the disease roll in the first place.
Our group just didn't like players dying because of a random bad roll without any chance to avoid it, even if it's "more realistic", whatever that means for a fantasy game.

[talysman]

I feel the percentages might not be correct mathematically, but I'm not sure. They definitely don't feel right gamewise, though. A 45 to 70% chance of sickness? That's kind of high. If one normal man attacks another normal man with a dagger, there's a 55% chance of a hit, and 50/50 chance that hit will kill... and that can be defended against with armor, or just somehow avoiding being attacked. Should you really be more likely to be felled by illness, including times when you aren't adventuring?

It's better if you reduce the number of rolls. Assume that characters rarely get sick during downtime unless there's an actual plague event or conditions are otherwise bad. Even then, a simple Con roll or Save vs. Poison once a season is more than enough. If you want to be gritty, maybe require a save every winter even in clean areas.

The other problem is that most illness is pretty mild, or at least something a stubborn adventure would just shrug off and keep going. More than half of all your illness should be colds or achy joints, something that is mainly a special effect. For my disease system, the primary effect of illness is that natural hit point recovery is slowed or stopped completely until the character recovers. Maybe, if the character is sneezing, add a roll to avoid sneezing while hiding. Maybe, if the aches and pains are severe, movement rates are halved. That's about all I usually need.

Severe illness, unless it's specific to a dungeon encounter and potentially avoidable, should be really rare. I think, from what you describe, you have an infection test (1 in 36 chances) followed by a disease type roll. I think that's basically the right course, but the serious illnesses should be really rare, and the potentially fatal ones should be rarest of all. And even the fatal ones should have a chance of not being fatal. If, say, there's a 1 in 36 chance of contracting an illness (per month or per season, normally) and a 1 in 36 chance that illness is potentially fatal, and a Con check (at half normal Con) to see if the victim recovers, plus some kind of non-magical healer available that could improve recovery chances, I think that would be about right.

[Scott Anderson]

The first question is this: do you really want to have a PC die from the flu or a stubbed toe that gets infected? It's quite realistic but it doesn't seem like any fun to me.

What do you think?

[hamurai]

Sept 7, 2017 17:47:37 GMT -6 Scott Anderson said:

The first question is this: do you really want to have a PC die from the flu or a stubbed toe that gets infected? It's quite realistic but it doesn't seem like any fun to me.

Exactly. This seems more in line of a game like Hârnmaster, which tries to emulate medieval harshness of life (at least that's how we experienced the 1st edition). D&D, no matter which version, is a fantasy game, so I wouldn't include random death by disease unless the player(s) made some really big mistakes.

[Scott Anderson]

That's a good take. I am really interested though, maybe people are interested in a subsystem for disease.

[foxroe]

Forgive me, it's been many years since my college statistics class, but I don't think that the percentages work that way. If you have an X% chance of winning the lottery each day, that chance does not improve as the days go by IMO - your chance is always X%, not X% times the number of days that have elapsed. So, the 5% chance per month is just that: roll d% every game month, and if the result is 01 to 05... <cough> <cough> <wheeze>. Otherwise, characters would be automatically diseased every year and 8 months...

And now to be hostile about the question

I just use saving throws whenever the characters find themselves in a situation where exposure is inevitable (i.e. rifling through the dead on a battlefield, swimming in sewers or fetid waters, Greco-Roman wrestling with an Acolyte of the Plague-Mother, etc.)

Edit: You could also build it into your encounter tables, with the likelihood of exposure increasing in cities and swamps.

[oakesspalding]

Great points.

So here's how I view the overall mechanic:

1. It's not to increase realism per se but, rather, to add fun decision making nodes for the players in terms of deciding whether or not to return to civilization, trying to find help from that wizard rumored to live in the valley, etc. (Granted, being infected by Plague may not seem fun, but still.) The Disease Table also includes positive outcomes - like gaining an ally or finding a treasure map, etc.

2. The random thing is only a subset of the overall structure and could be jettisoned. The primary uses for a disease mechanic are: A) To better explain or systematize what happens when you're bitten or touched by, say, a Giant Rat, Giant Mosquito, Giant Fly, Giant Tick, Giant Bat or rabid dog, Mummy or whatever. Though, again, such systemization is optional. You can always specify a unique disease for a particular monster. B) To explain what happens when a Cause Disease spell is cast. I envision the "normal" version to cause a random disease. I think that's more fun, though it may not be "realistic." A particularly evil or powerful Evil High Priest might be able to cause septicemic plague. Every. Time. C) To create fun and meaningful decisions for players where the odds are generally known but the outcomes (obviously) will not be. For example, should you enter that disease ravaged city to get help sussing out a magic item? Do you push on through that Malaria infested swamp or not? Should you spend time searching that room filled with dirty water? In my view, it's more fun if you usually have some idea of the odds. You "know" there's a 10% chance of catching dysentery or whatever. So does that make it worth it? It's ONLY 10%, after all...

3. Mild diseases (with one exception) are not represented. It's assumed that adventurers will always be grappling with hangnails, minor infections and the like but that's already built in to the system through the randomness of hit point determination, to hit rolls and so on.

4. The exception is the common cold, and it's exactly what talysman suggested. It's difficult to move quietly due to sneezing, etc. Plus it has a 5% chance of developing into pneumonia...

5. Except for Cure Disease spells or special circumstances, the characters would already be assumed to be doing the best they can to treat a particular disease, just as they are assumed to always do the best they can with treating injuries. Introducing, say, herbs and remedies into things in a systematic way would create a different game, though, again, there would be exceptions.

6. Each disease description has somewhat more explicit effects on play than talysman suggested (though not by much), with some diseases progressing through different stages involving different and unique symptoms, etc., which would render a character "weak," "very weak," "incapacitated" or, well, dead. Weak might be ST -2, CON -2, Attack -2. Very weak might be ST -5, CON -5, Attack -5, Recovery times often vary according to which stage was reached before cure, and you can halve it with complete rest, etc. For most diseases the character would be weak during the recovery period.       

  

[oakesspalding]

Sept 7, 2017 19:18:39 GMT -6 foxroe said:

Forgive me, it's been many years since my college statistics class, but I don't think that the percentages work that way. If you have an X% chance of winning the lottery each day, that chance does not improve as the days go by IMO - your chance is always X%, not X% times the number of days that have elapsed. So, the 5% chance per month is just that: roll d% every game month, and if the result is 01 to 05... <cough> <cough> <wheeze>. Otherwise, characters would be automatically diseased every year and 8 months...

Well, the "expected number" of your wins is indeed X% times the # of days. However, your chance of winning once (at least) is 1-((1-X%) * (1-X%) * (1-X%)...) for however many days you do it. For small chances (such as winning the lottery) the two results are approximately equivalent. Check my math, I just turned a year older yesterday so it may be faulty...

[foxroe]

Happy Birthday, oakesspalding !

Checking math...

[Scott Anderson]

While actuarial tables are rare, historical accounts that exist suggest that 3% of Rome's population died from communicable diseases between July and October each year and that this period was unusual. It is reasonable to guess - but it is just a guess - that the number of deaths from disease throughout the year was about 6%.  Rich people, about 5% of the population, fled the city during this period to spend time safely in the country.  

[oakesspalding]

Yes. And even as late as, say, 1800 in England, something like this might have been true. Life expectancy was low, even for the rich. Something had to have been killing people.

The question, though, is how this should apply to adventurers. They're Nietzschean Supermen, aren't they? Just kidding, but you know what I mean.

Cities were generally much more dangerous places for diseases in "normal" times. But my understanding from the history is when it came to the really bad pandemics, rural versus city didn't make much difference.

[Scott Anderson]

The average life expectancy is only part of the story though. 50% of Romans around 0 AD died before age ten. That means that when surviving childhood, life expectancy was more like 60. This seems right. A lot of body parts wear out by 60. Teeth specifically. Not being able to chew would put a damper on robust health.

But you're right. I was just looking at life expectancy charts in 1800-1850 and things in Europe were literally no better then than in Roman times. I bet there were good actuarial records in Britain and the US at that time that we could look at.

[Scott Anderson]

On the topic of how diseases affect you and what do you do to fight them, Arthur over on the Goblin Punch blog has tied that all up neatly here:  goblinpunch.blogspot.com/2016/06/the-glog-diseases.html

This is literally all you need to know about fantasy diseases and how to run them in a 5-page PDF.

[howandwhy99]

AD&D has an very elaborate system which doesn't specify the particular disease. I guess this is because there are so many potential diseases, but it also seems to remove the possibility of using disease like a poison attack. If disease is a possible weapon, I'd strongly suggest treating it like poison - With Poison Use proficiency and the Ferocity rules.

BTB AD&D uses different 2 checks:
1. Disease
a. (optional) Monthly check for everyone, Weekly if crowds, filth, hot and wet weather.
b. Spot check every time a known carrier exposes a PC. Monsters (people, animals, insects [giant?]) Items (food, drink, dirt, filth, etc)

2. Parasitic Invasion
a. (not optional) Monthly or weekly check as disease, slightly different termed conditions (?)
b. Spot checks, as above but slightly different exposure list.

Disease Contraction = 2% + Modifiers (lengthy table)
Parasite Contraction = 3% + Modifiers (lengthy table)

Roll for Body Area Affected (apparently only one per Disease/Disorder or Parasite)
- This determines the odds for the next 1-2 rolls

1. Occurrence (disease only)
- Acute or Chronic type. Chronic results can compound and lower later resistance odds.
2. Severity (both)
- Severity can be Mild, Severe, or Terminal

Modifiers to Disease Rolls Only (Table)
- Constitution score high or low, current disease or infestation, low hit points

Character Limitations upon contraction. These are cumulative. A limitation is gained per each body part afflicted AND per the severity rolled.

While I admit this is pretty thorough, I'd simplify myself. I'd really want to know the consequences of such odds to a city population before putting a system like the above in place. It also doesn't include how epidemics or plagues occur and how PCs can combat them.

[oakesspalding]

Yes. For any period with a significant infant mortality rate, it makes a big difference whether you include child deaths in the averages, and people often play fast and loose with this. Still, my impression is that life expectancies were still low in these examples, even when you take account for infant mortality.

Interestingly, even in recent or modern times, you can see how things progress. 37 years ago, Ronald Reagan's age was a big deal. He was 69 when he took office, a record at that time. Trump was 70 when he took office, a new record, but no one really remarked on it. Clinton was 69, and though some people claimed she had major health issues, it was framed as a Hillary thing, not an age thing. Bernie Sanders was 74. Rick Perry (who some people think acts as a child) is 66. Ron Paul is...but I digress.

But I think it shows that even in "modern" times and in "developed" societies, the perceptions and reality of aging can change relatively quickly. Did, rich people live past 40 in, say, 18th century England or the colonies? Yes, but not that many. And, of course, what complicates it is that for the people from those times that we know about, in many cases we know about them precisely because they bucked the odds.

Then again, there were many famous kings and statesmen in past times who died young. That is almost never the case now. I'm thinking of Baldwin IV (the Leper King) of Jerusalem. He died at 24 or 25, but still lived a full and important enough life to merit a Wikipedia entry 800+ years later.

His case also proves that leprosy wasn't as bad as we (or many of the people of the time) thought. Or you could gut it out, so to speak, at least for some years.

The leprosy was diagnosed when he was a boy. His father noticed that, in horsing around with other children, he felt no pain. After he became king he, according to one report, rode into battle with only one working arm. Some say his face was horribly scarred. Others say it was untouched and handsome.

In the 20th century he was reportedly a model for the French scouting movement. May he rest in peace.

[waysoftheearth]

Sept 7, 2017 19:36:07 GMT -6 oakesspalding said:

Sept 7, 2017 19:18:39 GMT -6 foxroe said:

Forgive me, it's been many years since my college statistics class, but I don't think that the percentages work that way. If you have an X% chance of winning the lottery each day, that chance does not improve as the days go by IMO - your chance is always X%, not X% times the number of days that have elapsed. So, the 5% chance per month is just that: roll d% every game month, and if the result is 01 to 05... <cough> <cough> <wheeze>. Otherwise, characters would be automatically diseased every year and 8 months...

Well, the "expected number" of your wins is indeed X% times the # of days.

*blink*.

So... are you suggesting that 5% chance of a win per day for 20 day yields 5% x20 = 100% probability of a win?

Moreover, are you suggesting that a 5% chance of a win each day for 100 days yields 5% x100 = 500% probability of a win??


In fact, the probability of at least one win in 20 days is 1 - the probability of no wins in 20 days.

If p(win) = 5%,
Then p(no win) = 95%.

p(no wins in 20 days) = p(no win)^20.
p(no wins in 20 days) = 0.95^20
p(no wins in 20 days) = 0.35848592

p(at least one win in 20 days) = 1 - p(no wins in 20 days)
p(at least one win in 20 days) = 1 - (0.95^20)
p(at least one win in 20 days) = 1 - 0.35848592
p(at least one win in 20 days) = 0.64151408

Therefore, where p(win) = 5% per day, there is a 64.15% chance of at least one win in 20 days.

Sept 7, 2017 19:36:07 GMT -6 oakesspalding said:

However, your chance of winning once (at least) is 1-((1-X%) * (1-X%) * (1-X%)...) for however many days you do it. For small chances (such as winning the lottery) the two results are approximately equivalent.

"Approximately equivalent" covers a lot of ground. But if you are suggesting that I have "approximately equivalent" chances of winning lotto with one draw as with 100 trillion draws, then so be it

[oakesspalding]

I'm not sure what you're arguing with. As for my first claim, I stand by it. To use your example, if you have a 5% chance of wining and you play 20 times, then, on average, you will have 1 win (5% x 20) at the end of the set. If you play 100 times, then, on average, you will have 5 wins (5% x 100) at the end of the set. That's pretty basic, unless I'm going insane.

As for your second claim or formula, it's precisely the same as my second claim or formula. Read it again. I used X%. You used p.

As for the two methods being "approximately equivalent" for small chances, I stand by that as well. If you have a 1 in a million chance of winning the lottery and you play 1000 times, then:

Your expected number of wins (or number of wins in the set, on average) at the end of the set will be .00010000.
Your chance of winning once (at least) will be .00009995.

Those two numbers are approximately equivalent.

[Scott Anderson]

You have approximately the same chance of winning the lottery whether you buy one ticket, or zero tickets. Practically the same chance.

[oakesspalding]

Sept 8, 2017 12:46:00 GMT -6 Scott Anderson said:

You have approximately the same chance of winning the lottery whether you buy one ticket, or zero tickets. Practically the same chance.

ARGHHHH!

1. That misses the point I was making above. To state the point another way, for very small chances and a number of tries equal to only a small fraction of the denominator of the chance, multiplying the chance by the number of tries gives you a very good approximation (off by less than 1% of the correct result) of the chance that at least one try will succeed.

2. Actually, your chance is infinitely higher if you buy one.

[oakesspalding]

So here are some tables. The idea is once you've determined that there MIGHT be a disease - by rolling that 01 or whatever - you choose the table that best applies. As you can see, there still is a healthy chance that the result will be "no effect." And even if a disease is indicated, whether you actually get it or not will depend on saving throws. Some diseases will generally only infect one party member. Others may have a good chance of infecting 50% or more.



Notes on the Tables:

“City” applies to city or any other crowded and potentially filthy conditions such as sieges and the like.

Use the Dungeon columns if the party spends the night in the dungeon.

Ergotism (actually a type of food poisoning) will only occur in inhabited areas.

In temperate climes, malaria will only occur in Marsh or Coastal regions.

Shipboard: Influenza will only occur in cold temperatures. Dysentery will only occur in hot temperatures. Malaria, sleeping sickness and yellow fever will only occur in the Tropics near a coast). After four weeks, any result of 01-25 will be scurvy.

Reinvigoration: All members of the party are cured of all hit point losses (even those at zero hit points). Though diseases, the effects of a level drain, unnatural aging and the like will not be cured. In addition, for each party member, hit points may be rerolled until a higher result is obtained.

Ally: The party encounters a friendly character or creature 1-3 levels above that of the most powerful party member. He, she or it will travel (and fight) with them for 1-6 weeks.

Treasure Map: The party finds a treasure map.

Magic Item: The party finds an unguarded magic item (though it may or not be benign).

[howandwhy99]

I've had magic items hang around like a disease. (Allies too). But I've never found a treasure map I couldn't drop on the spot. hmm?

[waysoftheearth]

Sept 8, 2017 8:08:42 GMT -6 oakesspalding said:

I'm not sure what you're arguing with. As for my first claim, I stand by it. To use your example, if you have a 5% chance of wining and you play 20 times, then, on average, you will have 1 win (5% x 20) at the end of the set. If you play 100 times, then, on average, you will have 5 wins (5% x 100) at the end of the set. That's pretty basic, unless I'm going insane.

I think I can see what you're trying to say Oakes, but the language is a bit confusing to me. You appear (to me) to be conflating two distinct concepts.

One thing is: the probability of a future outcome. If I play one set of 20 throws, then I will score a discrete integer number of wins and losses, which will lie somewhere on the frequency distribution of all possible results for 20 throws.

Another thing is: the summary of a collection of outcomes. If a number of people played the same set of 20 throws, each would get their own number of wins, per above. We could quantitatively summarise these results after the fact, and talk in terms of the mean, medium, mode, std deviation, and so on.


You were, I believe, saying that the second thing (i.e., "on average, you will have one win") should inform an expectation about the first thing (i.e., "the 'expected number' of your wins is indeed X% times the # of days").

I hope I can illustrate how the "average" result need not be the "expected" result.

So here's the frequency distribution for all possible outcomes of 20 throws with a 5% chance of each throw being a winner:



We can observe here (highlighted in yellow) that the most probable number of wins from any set of 20 throws is indeed exactly 1 win, and also that there is a 37.8% chance of this outcome. So what's significant is while one win is the most probably outcome, it is more probable (62.2%) that a player will not win exactly 1 game from 20 games.

Even though the mean and mode are close to 1 and exactly 1, respectively, it is not odds-on that I will get exactly 1 win from my next set of 20 throws.


This is more obvious in the case of 100 throws:



So yes, once again we can see that exactly 5 wins is the most probable outcome at 18.0%. However, observe that it is 82% likely that a player will not get exactly 5 wins from his next 100 throws.

So I guess what I'm highlighting here is that the most probable result need not be a very likely result, and need not be synonymous with the "expected" result.

A fine line perhaps... but hey.

[oakesspalding]

I agree with everything you just said.

Except for the conflation part.

I've never unjustly conflated anything in my life.

Re: my language. Interprete "on average you will have 5 wins" as "the expected value of your wins is 5.00." Of course if the number of throws is not divisible by 20, then the expected value of wins will not be a integer. At 85 throws, the expected value of your wins is 4.25." And so on.

The expected value is an average, or a mean, if you will, and thus has nothing to do with how many times THAT NUMBER might or might not actually come up. Obviously, when speaking of wins, any non-integer will NEVER come up, since one "win" is not really divisible. But the expected value is still an important concept - in many cases the most important concept - for all sorts of obvious reasons.

[oakesspalding]

Okay, I think the "No Effect" and beneficial results rows ended up making the whole thing too fussy, confusing and strange.

I was trying to build in the fact that diseases in general were much more common in certain climates/terrains/situations into the tables, but I think perhaps that's not the way to go. However, since Dysentery and the Common Cold are not that serious, you can approximate that idea while still having a disease for every die result.

Try these, instead (and mentally add a bit of spotted fever into the temperate wilderness mix):