Friday 31 July 2015

On Archery

When I presented missile range penalties for Explore, I gave a to-hit penalty of -1 per range increment over 20’ (where ranges go up as 20’, 30’, 40’, 60’ etc., doubling every other category). This fits the main overriding requirement - "simple" – but was unsatisfactory for a number of reasons which have niggled:

1)      Does this increase in difficulty reflect the real world?
2)      What about large or small targets?
3)      What about shooting long range against mass targets?
4)      Should there be a damage penalty for long range?

I left this for a while to ponder over, until I realised that I could resolve the first three together, the fourth I'll talk about another day.

The revised rules are:

For missile attacks there is a to-hit penalty of -3 per range increment over 30'.
Large or small creatures get a penalty versus missile attacks equal to their size.
Missile attacks versus mass targets get -6 at any range.
If you care which of the mass targets is hit, roll to choose a target, then roll to hit it.

The details of how I derived these rules follows, but the derivation itself is of less importance than the fact that they are derived from observations, not purely arbitrary, which is a core design philosophy for Explore. When you design this way you find that the rules work together, whereas with independent rulings they often don't mesh well.

To-Hit Penalties For Range
Following Delta's post on archery accuracy, I agree that it is best to model increasing the range to a target by instead reducing the size of the target. That is, if you double the range of a target then that's the same as making the target one half the diameter. Also I agree you can model the distribution of shots as a bivariate normal distribution (i.e. both the x and y positions of the shot are normally distributed). Adjusting the figures to match the data he cites for "Britain's finest archers" I get the same answers as his calculations. For the range categories in Explore you get:

Target #







I have extended the list past normal ranges as those values will apply at shorter range for lesser archers, added the "simplified" range category, and the equivalent target numbers and chances with 2d10 (open).
Each increased range category equates to halving the area of the target,  and as such you would expect the ratio of successive chances to tend towards 2. You can see this at the bottom end of the results (10%, 4.9%, 2.5% etc). This matches exactly adding three onto the target number in Explore.
On the right half we give the new rule for Explore (each increase in range gives -3). We can get an exact match for the start of the results, or the end of the results, but not both together. I have opted for the low range results to be accurate, with the long range ones being half the predicted, with the hand-waving explanation that long range presents extra difficulties in judging distance.

Point Blank Range
In Explore "finest archer" would be +13 (+3 from stat, +10 rank bonus). The lowest parry you can get is 4 (10 -3 for stat -3 for surprise), so that's the parry for a stationary target. So the standard target # they need would be -9, hence they're getting 11 penalty for range, so roughly four range categories. So this would make point blank to be 30'. This is a slight increase from the previous 20', but as range penalties are now quite severe that seems fair.

Large and Small Targets
For large targets, +5 in size makes you double the height and six times the weight, so 2 x 1.73 x 1.73 times bigger so 3.46 times the cross-section. Hence for every +3 in size you’re roughly double the cross section, so present the same size target as being one range step nearer, so should get -3 versus missiles. Thus you should get a penalty versus missiles equal to your size (which is a +3 bonus versus missiles for halflings).

Shots versus Mass Targets
Consider shooting at a mass of targets compared to a target placed 30' away. A single person 40' away would fill half that target. Four people 80' away would also fill half that target, and even 512 people at 960' away. Hence each group of people is the same difficulty to hit - it's the density of the troops that matters. Since troops are unlikely to completely fill the target, I'm going with -6 versus massed troops (instead of -3).

In summary, following Delta's observations about real world archery gives us a single principle that gives us rules for three different situations.

Friday 24 July 2015

Please read this if you're interested in justice

Please read Raven Crowking's blog and help out by signing his petition for justice in the case of his nephew, Wally. It reminds me of a recent tragic case at my children's school, in this case the child was posthumously cleared.

Thursday 16 July 2015

The Worst:Trollstone Caverns

My First RPG
My first RPG was Tunnels & Trolls (Fifth Edition) in 1982. We only played it a few times before moving to Dungeons & Dragons, but I always retained a soft spot for T&T, especially Liz Danforth's fantastic art.

I knew it was more famous for solo play than for groups with a DM, but I have always assumed this was due to a distaste for the combined combat outcome - that is, whichever side wins each round takes no damage, the other side all take damage (lessened by armour).

The combat system, in brief, is: everyone makes a combat roll, you take the total for both sides, the losing side takes the difference as damage (distributed evenly) but reducing this by each target's armour.

The Kobayashi Maru
A couple of years back my son wanted to try his hand at DMing so I suggested T&T, and he ran us through the introductory adventure Trollstone Caverns. In the very first combat, which is apparently unavoidable and is evidently to introduce you to how combat works, we had to run away before we were all killed - and it appeared almost impossible for us to win the battle, in fact it appeared nearly impossible to even wound our assailants.

I thought my son must have got something wrong about the adventure, so later I looked at the encounter. You get one monster per PC in a one-on-one combat, of a random type. For example, the weakest monster has MR 16 which means it gets 2d6 + 8, the strongest is MR 24 with 7 armour which means it gets 3d6 + 12. If I tell you that the sample character included in the rulebook, Fang the Delectable, has 3d6 - 1 for combat, you can see he's in trouble against even the weakest of the possible monsters: he averages a combat roll of 9.5, while the weakest monster averages 15, even with 6 armour he's taking damage half the rounds, and is exceedingly unlikely ever to score a hit. Against the other monsters he's dead without a chance. And one monster is immune to non-magical attacks.

I just ran four sample combats of Fang against the weakest monster, and he died four times in succession (I would have run more, but they went on for ages). He only did 3 points of damage in total over all four battles. So the sample character in the rulebook is almost guaranteed to die in the first combat.

Now our characters (digging out the character sheets and the session notes) had:
  • Viper (Rogue): 4d6+7, 8 CON, 8 armour
  • Hawk (Warrior): 3d6+3, 6 CON, 24 armour
  • Flash (Wizard): 2d6 -1, 7 CON, 8 armour
Which is better than Fang, but my son rolled two Wargs (MR 28) and a Shadow Ghost (MR 12, vulnerable only to magic). It was only because we'd all played RPGs before that we knew to run away after a couple of rounds - if this had been our first experience of an RPG we'd likely never have played an RPG again.

So, the introductory adventure for new players to RPGs includes an unavoidable first combat, which has a high likelihood of being unwinnable!

(Evidently, like Kirk, we should have stolen the module the night before, and altered it to guarantee victory).

Digging Deeper
I initially put this down to a horrific lapse in playtesting this adventure, but this got me thinking about the combat system, so I ran a computer simulation. What I discovered was that given a particular character’s dice and adds there is a very narrow range of MRs between those monsters which stand almost zero chance of winning and those against which you stand almost zero chance of winning.

I ran a sample combat, and found that below an MR of 19 the monsters had less than a 1% chance of winning against my PC, and above an MR of 24 my PC had less than a 1% of winning.

If you have group combats against multiple monsters, combining the results from both sides decreases the chance of a freak result - the outcome of the combats become more and more certain.

For example, I ran five PCs versus one monster, and found that the below an MR of 110 the monster had less than a 1% chance of winning, and above an MR of 118 the PCs had less than a 1% chance of winning.

In addition, the average length of a combat went from a well balanced match giving an average of 9 rounds with one PC, to an average of 70 (!!) rounds with five PCs. Note that this is with minimal armour (5) - extra armour makes the combats yet longer: with 5 PCs and 10 armour it goes up to 1900 rounds. I think we call that a stalemate.

So with a party of characters, almost every combat is a fait accompli, and the few that aren't go on so long they're effectively a stalemate. I can't believe this system was ever played with parties of PCs as written.

A Saving Grace?
Any reference to T&T combat talks about the importance of Saving Throws in combat, but in Fifth edition this is only mentioned in the section on "Unusual Combat Situations", and the only example is a simple "I can't win this fight, so I try a daring do-or-die manoeuvre." If this fails you take all the monsters hits as damage (so you're likely dead), if you succeed then the monster takes all hits as damage (and if not dead, its MR is so reduced that you'll win). This is a "last throw of the dice" final desperate tactic - not an expected feature of combats.

Calling All Stations!
So has anyone actually any experience of a long running T&T multi-player campaign with the rules as written?

Is it really only suitable for solo play?

Do the combat manoeuvres and "spite damage" in later editions fix things?

Or is the excitement over the Deluxe Edition Kickstarter purely about nostalgia and love for Liz Danforth's fantastic art?

If you love the game and think I'm being truly mean, redress the balance and show me how I'm doing it wrong. Until then, I'll just look at the art.

Monday 13 July 2015

The Best: Hand Drawn Maps

Whilst thinking about Geomorphs I was reminded that Dyson mentioned that his style was reminiscent of Chaosium's hand drawn maps in early Runequest modules. 
...why make the maps pretty?
Because it adds to the GM immersion into the game. Gaming is about roleplaying, and anything that gets you further into the mood and setting is a major boon. So old-school styled maps like the ones I draw remind me of the old Chaosium products that I loved when I was a young one in the hobby - and thus they get me in the mood for that style of game again, thus helping me run such a game.
Now I haven't seen much of the old Chaosium products so I can't comment on their maps, but for me, hand drawn maps remind me of adventures in fanzines, such as this one featuring cross-hatching from Demons Drawl #9 Dec 1984:

Lawrence Beckett has done a very nice drawing of the abbey with matching floorplans. Seeing this again triggered a memory of the old Pelinore maps from Imagine Magazine - these by "Kh" from August 84 look like the inspiration for Lawrence:

Side elevations, floorplans, and cross hatching. I particularly appreciate the locations named on the map in lieu of a key.

For evocative hand drawn dungeons I was always drawn to the dungeon levels in City State of the Invincible Overlord. I liked the odd shapes and the odd noises - laughter, moaning, growling...

My own early efforts at the time were sadly a little lacking... but this effort from 1983 (when I was 11) is interesting to me as it's a one-and-a-half-page dungeon (the other half page is on the back):
Don't ask me why I used pink paper! Notice the key fact - there is no key. My earlier efforts had a keyed map, by this time I'd already realised it was better to put the details on the map. In fact many of my old TSR modules have pencil marks on the maps where I'd written directly on the map what was in each rooms.

I see many attempts by people to produce "professional" looking maps using computers, and I look at these hand drawn maps, and I prefer them in every way. It's not just nostalgia - they're evocative, unbounded and free. When you draw free hand you are limited only by your imagination – notes, illustrations, side-elevations, isomorphic, multi-levelled. Anything that occurs to you can be added.

As an aside, my Geomorph investigations also stumbled across "Pythagorean Tiling" which can be used to tessellate 10x10 Geomorphs and 15x15 Geomorphs into a pattern which obscures the joins between tiles by staggering them:

This looks quite appealing, and stops your dungeon from being rectangular. I'm still considering the pros and cons of having edge pieces (which this pattern makes more difficult) or removing any passages which go off the map.

Sunday 5 July 2015

Geomorphs: Resurrecting an old idea

I first encountered Geomorphs with the 1978 boardgame “The Sorcerer’s Cave”. I can’t recall that much about the game itself, but the vision of a cave system formed randomly be the placement of tiles always intrigued me – though the realisation of them in this game is very simplistic.

I don’t think I saw anything about Geomorphs again until resurgence of interest about them in the last few years on the internet.

In 2013 the PCs in my D&D game needed to destroy an artefact, so set off in search of “The Juggernaut of the Eternal Labyrinth”. I'm not about to design an infinite labyrinth, and pure mazes are very boring, and my experience of random dungeon generation has been that they’re not that good, so I settled on a Geomorph solution. Simple Geomorphs with a single exit on each side provided a suitable “labyrinth” feel. I made nine Geomorphs and laid them out randomly on the printer/scanner, printed them out, trimmed to size to form a larger Geomorph and repeated:

Combined with a random encounter table, and suitable appearance of the Juggernaut it worked very nicely.

Gary Gygax appears to have invented Geomorphs in 1976, but the concept didn't appear to take off. I'm not sure when Geomorphs resurfaced - there is a 1998 TSR product TSR9556 “Dungeon Builder's Guidebook” with 78 Geomorphs but they seem to have lost the plot completely and don't even seem to fit together. In 2004 Goodman Games released “Dungeon Crawl Classics #9: Dungeon Geomorphs” which are proper Geomorphs, quite large (14 x 19) with 2 connectors on each side.

In 2009 Dyson started his own Geomorph challenge using 10 x 10 Geomorphs with 2 exits on each side. I'm not sure if he originated this format, but this became the de facto standard. Because they're square they can be rotated, and reusing this standard enables Geomorphs from different authors to be used interchangeably.

When my son started DMing OD&D this year he was intrigued by the Gygax originals, and I was surprised to find that they were quite different to the modern concept. 

Like todays standard, the originals are square so can be used "in any position (not necessarily "right-side up")", but they are larger (21 x 21) and have three connector points on each side, but one of them is typically missing. This means that there is always one connection between Geomorphs, but in a varied position.

I got to wondering about resurrecting this concept, then realised that a 15 x 15 Geomorph with 3 connectors on each side could work. If you lay out 4 of these as a square, they would match a 3 x 3 square of  the standard variety, hence would be compatible. With the larger size, you could place chambers or other features off centre. Furthermore, in a larger size there is space for dead ends to lead to features.

So the template would be 15 x 15 square, with 3 connectors at the 3, 8, and 13 columns and rows. To start you off, roll d4 for each side to see if the dead end is left / centre / right / no dead end. If you get a dead end in two adjacent corners, you join them together. When you go down a connector, 3/4 of the time it will go through, 1/4 of the time it will be a dead end.

So the plan would be to mix and match the two types of Geomorph, say put in a vertical or horizontal strip or an embedded section of the other sort. This would break up the repeated pattern of the connectors.

As a proof of concept I've done 12 quick 1" square Geomorphs with a felt tip pen, scanned and inverted them, and then plugged them into a bit of javascript to lay them out randomly. If this page gives you a "secure content warning" that's it complaining about the javascript. Reloading the page will give you a different version of the random Geomorph dungeon below.

I like them - the dead ends seem to work OK, and the repeated pattern is obscured - so I think I'll do some proper ones. One thing I've learned is - if it took that long to create sketch Geomorphs, I think it'll take a bit of time to create 12 proper ones!