Showing posts with label math. Show all posts
Showing posts with label math. Show all posts

Sunday, April 2, 2017

Space Gaming - 2d vs 3d maps

A discussion group on Facebook that talks about Traveller related items has been talking about 3d mapping vs 2d mapping.  I am a big fan of Classic Traveller (1977) and its derivative games.  Those include MegaTraveller, T4, T20, T5 and Mongoose Traveller.  One of the things about them, however, is that going back to the original, they render star maps in a 2d fashion, typically on a hex grid.  Let me say, that I am a big fan of the 2d representation, because it is not true (form a Physics sense) but it does represent a presentation of game/setting information in an intuitive and extremely useful format. What could be better?

On a Traveller map, the first thing you will notice is hexagons, some with "world systems" marked in it.  Each hexagon represents a parsec (roughly 3.26 light years), and any parsec that has a solar system in it that is potentially of interest, has a marker on it showing such.  This is a nice reality on the space involved, for each hexagon might be home to several solar systems (if a denser region of the galaxy, this might be a higher number), or it might be void.  But regardless, only "interesting" systems are marked.  That means, systems that either might have a civilization/colony on them, or might represent the possibility for such.

Regina Subsector - from the Traveller Wiki

But the maps are in 2 dimensions.  Originally they were in simple black and white, and these days it is common to find them in color, with the various regions and system colors indicating some details.  But the maps are still on a 2d hex grid.  So on the example above, the planetary system Forboldn is 1 hexagon away from Knorbes, which means an approximate 1 parsec difference between them.  This is great for the game, because the interstellar capabilities of the ships are measured in the ability to 'jump' between points of normal space, that are limited in (because of the technology involved) distances of roughly 1 parsec.  So a jump-1 ship could travel from 1 hexagon to another (a vast distance) in the space of a week - but that makes this rendering of the starmap very useful for game navigation.

But what about a 3d version of space - you know since space is basically a 3 dimensional construct, and all that...  Well, there have been some game maps that have done a good job on this.  I would like to talk about three, briefly, but in the light of a tabletop gaming constraint.  That constraint is that the map should be able to be printed out and passed around at the table, as an artifact to be consulted during game play.  A very nice 3d map on a computer that can be rotated, zoomed in and out of, is fabulous, but of more limited use during tabletop gaming.  In computer gaming, however- but that is a different story.  On to the examples I have.

Space Opera
In the game Space Opera, the materials that were released by Fantasy Games Unlimited (the publisher of the game) were often placed on planets in their published Sectors.  These were regions of space, published as game setting modules, that featured a number of planetary systems in a cube of space.  The examples below are from The Mercantile League (which, if I recall, is Star Sector Atlas #2).

The maps were published in their 8.5x11 Star Sector Atlas books, and the scale (which required the use of a ruler on the page to determine distances) was typically 1mm to 1 light year.  So you get out your ruler, determine that two worlds are 36mm apart from each other on the page, and you say "They are 36 light years apart".   Without having to use X and Y coordinates, and without having to do the Pythagorean Theorem calculation -- square root of ((x1 - x2)^2 + (y1 - y2)^2).


But, the 3d portion comes in with the fact that each world had a positive or negative measure (in light years, or LY) above or below the plane of the map.  So, taking the difference of these two gives you a second measure, the difference in the Z coordinate between the two points.  So now, with the X/Y measure in hand (the distance on the page), and the Z measure (calculated from +/- height differences) you can get the absolute 3d difference by again applying pythagoras as above.

The game publications gave you, in each Star Sector Atlas, a basic travel distance table that listed the distances (and some game economic factors) for the pairings of the more interesting/important worlds within the sector, so the distances were already calculated.
Finally, the last thing that was provided, in terms of a mapping assistant, for the FGU Space Opera Star Sector Atlas products, was the map showing the common space lanes (i.e. - the ones described in the "Fares and Cargo Rates" table).  This was simply the sector map with star lanes drawn on it.

Pros/Cons of the FGU method - these are easy enough to understand, and having the Z coordinate on the map makes for an easy transition to 3d.  As always, with a 2d depiction of 3d space, the map can be illusionary in some instances, where there is a sharp Z difference between two points that are otherwise (X/Y) close on the map.  These maps are nice because there is no real grid (they would come, later, with an Index, matricing a number vs. a letter axis, to make finding planets a little easier), but the obverse of that coin is that you need a ruler graduated in millimeters (or would have to convert inches, each to 25.4 mm), in order to find distances that are not on the quick look up table.

Universe
The SPI effort to enter the roleplaying game market was with the game Universe.  Universe had three very nice things going for it.

  1. The character generation sequence had a great way to compare physical stats for characters from different geophysical planetary backgrounds (i.e. - difference in physical parameters based on the planetary environment you were born/raised in). 
  2. The starship construction/combat rules were very nice (describing a setting with generally common ship hulls, perhaps as in the Niven Ringworld universe, but with different modular components), as you might expect from an experience board game publisher that had devised printed several sets of starship combat rules by the time the RPG arrived.  In this case it is the DeltaVee game.
  3. The 3d space map of the area roughly in a sphere of about +/- 25 light years around Earth.

This article is interested in discussing the space map.  It worked, in concept, very similar to the map from FGU but with some important differences.  It incorporated iconography and color (as do the more modern Traveller maps) to indicate something about the stellar systems mapped.  In the case of Universe the colors correspond to the stellar sequence.  Different from the FGU Space Opera maps, it used a graph/grid to show the location of the worlds, so you did not need to use a ruler to find the difference of planets. But it did mean you had to apply Pythagoras.  One of the nicest things about the Universe map is the fact that it included the X,Y and Z coordinates for each system.  Also, Earth/Sol is at (0,0,0) so measure to other worlds from Earth is pretty easy to do.


Finally, the third method of showing a 3d space on a 2d map is from the old Metagaming science fiction empire building game, Godsfire.  This game represented three dimensional space on a hex map.  The measurements for space were abstracted (slightly) and objects pressed on to a hex map, to make counting and measuring movement and disances between star systems to be quite easy (just count the hexes).  The Z component was brought in by making sure that each hexagon only represented the space at the plane of the map.

Within the hexagon, there was a series of 11 spaces, 5 representing levels below the map, 1 representing the planar level of the map, and five representing levels above the map.  As this was a map for a multiplayer wargame, moving objects on the map was necessary, and could be accommodated by moving the object (ship, fleet, etc) by moving it "up" or "down" within a single hex, or from hex to hex (arriving at the same level in the new hex, as the level that you departed the old hex from.

Pros/Cons - It is abstracted (as mentioned), and not as precise as the other methods, but it is certainly quick to navigate and move pieces around in.  Very simple and straightforward for campaigning and wargaming.
As you can see from the image, each "hexagon" is actually represented as a square, but using offset rows, means each is spatially related to the six around it (essentially a hexagon).  And the spiral showing the levels (from +5 down to -5) within the hex.  Very convenient.

This map was used by a group I played in years ago, as the campaign map for a Leviathan wargaming campaign.  It was an elegant way to do three dimensions, and it was very interesting to see players get surprised by a fleet that appeared similar, only to realize that it was at +4, and they were -3, adding (effectively) 7 more spaces of difference between the two fleets.

So, there.  Three different methods of doing 3d mapping, on a 2d space, with different pros and cons for each.

If interested in this topic, I can (not strongly enough) recommend Winchell Chung's page on 3-D Starmaps over at his Project Rho website.  Winch is a wargamer and artist and generally interested in the intersection between science and sci-fi and gaming, and his web pages are chock full of the most excellent information.

Friday, March 24, 2017

Dice Spreads, Probability Curves, and Rules Writing

In advance of an upcoming article I am planning to write, where I review the Knights and Magick rules (by Arnold Hendrick, and published by Heritage back in 1980), I wanted to address a rather interesting dice mechanism that are employed in the rules.


That is the 1-15 dice roll.  This is accomplished by rolling two dice, and adding them together.  The first dice is a rather common d6, and the second dice is the typical d10, but numbered from 0-9 (the 0 face on the die is valued at zero, and not at ten).  Adding the two results together yields a number from 1 to 15.

Now, the asymmetry of the roll always bothered me.  How could you possible hope to have game feng shui when you are rolling two different dice, and adding them together?  It was preposterous!  Sort of like the damage rolls from early Runequest and Call of Cthulhu and other games, where you might have 1d8 for the weapon, plus 1d4 for your strength, and maybe +1 for a good quality blade - so your damage looked like 1d8 + 1d4 + 1.  Weird, but okay for roleplaying games.  Wargames were supposed to be above all that sort of thing.  Better.  More pure.

But here we were, with Mr. Hendrick's rules using a D6 and a D10 added together.

Let's look at the results.  To compare it, a brief examination of the (much more common) 2d6 bell curve.  Here we see a smooth progression from one result (for a Two), up to six results for a Seven.  And then back down again to one result for a Twelve.  There are a total of 36 possible dice pairings, so the frequency is a number of times out of 36.  This is pretty standard stuff, that is part of any study of probability, but also should be pretty intuitive to just about any long time gamer, and/or game designer.

ResultFrequency
21/36
32/36
43/36
54/36
65/36
76/36
85/36
94/36
103/36
112/36
121/36

But now lets take a look at the rather interesting D15 result.  Here we have 60 possible results, rather than 36.  There are more possible results (1-15, rather than 2-12).  Plotting the possible results we see:

ResultFrequency
11/60
22/60
33/60
44/60
55/60
66/60
76/60
86/60
96/60
106/60
115/60
124/60
133/60
142/60
151/60

Okay, so what do we see from this?  First, the numbers of all the middle results, from 6-10, have the same frequency - 6, our of 60 results, or a 10% chance of getting any of those numbers.

How does this behave in the game?  Well, we end up with many more middling values (half the possible results are from 6-10).  Plus, in a game that deals in + and - factors added to the dice, for a variety of different causes, this will tend to balance out and level the impact of the dice modifiers.  What I mean by that, is that in a 2d6 dice roll, a single +1 or -1 can have a very high percentage impact on the dice chance, especially if your base number is off the middle.  In this case, the middle (or stable region) is spread out, so that dice modifiers are more predictable in their impact (a minor change with a +1 or -1).  This also means that having additional plus or minus factors won't be an overwhelming impact, as it is in a 2d6 curve.

Once I pulled out the results, and looked at them, the asymmetry doesn't bother me AS MUCH, but it is still there.  I think I would prefer a 3d6 roll, or 5d4 roll, to balance out that middle and level it - but that is a different story.

Now, with the recently reviewed Hackbutt & Pike rules (written by Ben King, in the Tac-50 rules series), we see that the casualty table is driven by a dice difference.  Again, looking at the 2d6 probability table, but selecting results based on the difference between the two numbers (agreeing that doubles results have to be rerolled), we have the following:

Die DifferenceFrequency
110/36
28/36
36/36
44/36
52/36

Of course, the total frequency of pairs with a difference only adds up to 30 out of 36, again because we tossed out the possible results where doubles are rolled on the dice (6 chances out of 36).

This is a very flat progression.  It is also interesting that on the Tac-50 tables, the least valued result is always a difference of 5, followed (in order) by 1,2,3,4.  A very interesting use of an interesting probability spread.



One more thing - I have a plan to (very soon) write a review of the original Sword and the Flame rules.  In those rules, as in many others before and since, combat is resolved between two miniatures by rolling a dice on each side, and the high scorer wins the fight.  With no modifiers, and with Tie results not counting, there are 15/36 chances for each side to win, if using six sided dice.
  • Side A wins - 15/36
  • Side B wins - 15/36
  • Neither Side (tie) - 6/36
But what happens when we change this by just a little bit?  Say, as in The Sword and the Flame, one side or the other gains a +1.  In this case, what if Side A gains a plus 1?  Very different results.
  • Side A wins - 21/36
  • Side B wins - 10/36
  • Neither Side (tie) - 5/36
Looking at the results it shifts, from even chances, to a Two-One chance of winning - from just granting a +1 on the dice.

With a +2 on the dice, it is even more extreme.  In fact, looking that the following table, we see it is 5-1 in favor of the side with a +2 on the dice.
  • Side A wins  - 26/36
  • Side B wins - 6/36
  • Neither Side (tie) - 4/36
Now, lets stretch this even further, and take a look at what happens, with no bonuses, but with one side winning Ties...  Again, lets assume that Side A has the advantage, and with no dice modifiers, will win ties.
  • Side A wins - 21/36
  • Side B wins - 15/36
  • Neither Side (tie) - No Ties
From these results, the odds change slightly, from even chances, to 4-3 chances for the side with the advantage.  Not as extreme as a +1, but then again, conceptually, saying that one side or the other wins ties is like saying they got a +1/2.

What we see is that in the case of an opposed dice roll (found in games going back to Featherstone and Grant, of course, but still present in modern sets) getting a +1 or even the benefit of winning ties is a very large bonus.

This idea of opposed dice rolls is common in many rulesets.  But in some others, there is the possibility of a player rolling several dice (typically more dice, for higher skilled combatants, for instance), and then selecting the highest dice, before comparing it to the opposition.  In this case, for instance, there is a much greater chance for a soldier who rolls 3d6 and selects the highest, to have a great number than his opponent, who is only rolling 2d6 to select the highest.  The component of a good number is still present, as a singleton dice can always come up as a 6, and the highest opposition roll could be any value less.  But what is the probability?

Lets construct a table, where we record the odds (and percentage chance) of getting a number, between 1 and 6, if it is the highest (or tied for the highest) out of a pool of dice.  To keep this simple, and illustrative, I am going to do it for three pools of dice - a single D6, 2d6, and 3d6.  Notice the shift in probability...

Highest
Result
1d6 Prob.2d6 Prob.3d6 Prob.
11/6 (17%)1/36 (3%)1/216 (.5%)
21/6 (17%)3/36 (8%)7/216 (3%)
31/6 (17%)5/36 (14%)19/216 (9%)
41/6 (17%)7/36 (19%)37/216 (17%)
51/6 (17%)9/36 (25%)61/216 (28%)
61/6 (17%)11/36 (31%)91/216 (42%)

Again, there is a dramatic shift in probability, just by changing the dice rolling mechanism slightly.  By rolling 2d6, and selecting the highest number, your chance of having a 2 is halved, and your chance of having only a 1 is dropped to approximately 3% chance.

So the chances of rolling a higher number, when you can pick from a pool, are better.  That is intuitive, but looking at the table for 3d6, we see over a 40% chance that your number will be a 6.   And if you need AT LEAST a 5 (which means that the results for a 5 or a 6 will satisfy) your chance increases to 60% (42+28), and so on.

At this point, I will abandon the exercise.  It might be nice, to compare the chances of a player rolling a pool of 2d6 to beat a player rolling a pool of 3d6, but I think I would rather return to writing reviews.  And maybe lunch.