2013-09-28

SciFi Saturday – Point-Buy System

The Star Frontiers: Knight Hawks game comes with a small number (4) of pre-designed scenarios, but no guidelines for constructing or balancing your own games after that. While this was common for the era (it's the same approach as in Doug Niles' other TSR wargame of the era, Battlesystem for D&D), I now think it's a serious oversight. I wanted to create a point-buy system to make pickup games of SFKH a snap, and my first thought was to use the spaceship pricing from the Campaign Book for that purpose (see last week). It became pretty obvious that was a total failure -- the small ships just always wreck the large ships at the prices indicated. But if we massage the prices a bit we can get something more reasonable -- namely compress the prices closer together (increase cost for the small ships, decrease cost for the big ships). Here's what I've come up with (edited costs in "mega-credits"):
 
Ship Cost
Scout 2
Frigate 3
Destroyer 4
Light Cruiser 5
Heavy Cruiser 6
Battleship 8
Assault Carrier * 10

* Assault Carriers include 6 Fighters each.

As always, my doctrine is that "The acid test is gameplay". You just can't balance this stuff by looking at any a priori formula on the statistics, as the interactions are complex enough that you simply have to see how they actually fight against each other. And frankly that's hard. What I used to assess this is to take the earlier price-balancing simulator for my Book of War game, rip out the guts and rebuild it to handle SFKH spaceship combat, and then run a few tens of millions of mock combats to see what happens. The result is that the best combination of prices (keeping it near the campaign book pricing) is what I came up with above. Here's some analysis as spit out by the simulator:



Assessed win percents (budget 10-30):

.. VS FF DD CL CA BB CV W AvgPct
-------------------------------
VS -- 55 88 77 82 67 -- 5 67
FF -- -- 95 88 90 79 59 5 76
DD -- -- -- 51 58 -- -- 2 30
CL -- -- -- -- 72 -- -- 1 38
CA -- -- -- -- -- -- -- 0 24
BB -- -- 63 58 78 -- -- 3 49
CV 66 -- 82 66 74 56 -- 5 64




What this shows is, for the same total budget, the percentage of time that a fleet of one type of ship beats another type of ship. For example: Assault Scouts (VS) beat Frigates (FF) 55% of the time in a straight-up fight with the same cost on both sides. Assault Scouts are favored against 5 opposing types (under "W" for wins), and their average win percentage is 67% across all possible enemies.

So the resulting game isn't perfect by any means. There are clearly 3 preferred types, namely Assault Scouts, Frigates, and Carriers (VS, FF, CV). The other large capital ships are pretty much also-rans. But after trying many combinations it's the best I could come up with, and there's really isn't any wiggle room left in the prices to massage it any further (notice the sequential prices 2, 3, 4, 5, 6 all used above). This assumes that we keep the ship statistics as shown in the book (although scouts get 25 hull points), and we keep the costs simple with small integer values (and near to the book pricing results).

There is the nice fact the there isn't any single best type, and we have at least one rock-scissors-paper cycle to keep the game from being totally solved. Notice that Scouts usually beat Frigates (55%), Frigates beat Carriers (59%), and Carriers beat Scouts (66%). That itself was not easy to make happen. Battleships aren't terrible, either (arguably their price could be 9, but then you get a pattern where every ship beats everything bigger than itself, and battleships lose to everyone).


The simulator is fairly crude, but based on the tightness of the price-points available, I don't see changes making much difference to the ultimate result. Here are some assumptions built into the simulator: (1) It only handles the Basic Game (not the Advanced Game with hit-location, damaged systems, etc.). (2) It doesn't deal with any movement -- ships are simply set up 3 spaces away and blast away at each other. (3) Faster-moving ships are given first offensive attacks (what I see effectively happening in table play). (4) Forward-firing weapons don't get defensive shots (as I see in table play; change that and capital ships fare better, esp. frigates). (5) As soon as an enemy ship blows up, the attacker can switch shots to another (technically not the game rule, but it approximates some intelligent prior shot selection by the attacker). (6) As noted above, Assault Carriers carry 6 Fighters into the game with them (same as what we see in the published scenarios and campaign order of battle).

The combat simulation runs on a random budget of 10-30 mega-credits (again, about what we see for the published scenarios in the Tactical Book). Any remainder is spent on a pro-rated "partial" ship (damaged?) so we don't get weird artifacts like Carriers losing out just because the budget was 28 (and thus simply flushing the last 8 points of value down the drain). Interestingly, the results change for different budget levels. Like, at 10 mega-credits, Frigates beat everything (while Battleships beat everything else); but at 15+ mega-credits, Carriers beat everything (and Assault Scouts win out over everything else). I might recommend a game at 12 mega-credits, where the matchups seem to be as shown in the table and graphic above. Future research: Combinations of ships, tactical movement, and use by intelligent players may modify any of these observations, but I think the point-values given are a solid basis.

If you did use these as the revised prices for warships (in mega-credits), then you have to think about where the Campaign Book system gets tweaked. Maybe engine costs go up for the small types (add a separate expense for hyperspace drives?), while reducing number of engines for the bigger types or something. While the cost for the Heavy Cruiser and Battleship went down (reflecting their relatively low utility, hit points, etc. for their price), we note that the Assault Carrier price is actually pretty much what we computed last week from the Campaign Book (7 for the CV + 2 for the VF at cost 1/3 each). This reinforces a place where I think the rules have a big gap -- there's no line-item or expense for the fighter hangar/launchpad/recovery system on the Carrier. Perhaps if such a system had a cost on the order 3 million credits, and the total was rebated as per the Battleship type, then that would explain things.

Minelayers don't fit into the simulator at all (nor can I see how to use them in my tabletop miniature games), but by interpolation we can guess they would cost 3 mega-credits (same as the Frigates). If you'd like a convenient index-card sized handout to pass around your table, then that's linked below.


Okay, so there's some of the things I was thinking about on this exploration. Finally, if you want to see the Java code yourself and tinker with the settings and prices, then that's the last link underneath here. (Note: Work in progress; next code release due next week.)



3 comments:

  1. One way to further tweak prices without using decimals or fractions is to simply double the point pool and all the point costs. That way you have twice the resolution, if you will, when trying to fine-tune prices.

    ReplyDelete
    Replies
    1. You're right, of course, but I wanted to lean towards keeping things in the same range as the official pricings from last week (and I don't think that the double-granularity would be enough to make a big difference).

      Delete
    2. So I just did that and it makes more difference than I thought: suggested costs are then 4, 6, 7, 9, 11, 16, 18, and all the average win percents are within 50%+/-7%. But the Light Cruiser is then the best type, and one can debate whether it should be price 9 (thus the best) or 10 (and thus the worst, beaten by all others and 37% avg. win percent).

      Delete