HG Starships - Professor Lenat and EURISKO’s Winning Fleet
By: Nick Munn <email@example.com>
Mark Clark <firstname.lastname@example.org>
Django Upton <DUpton@vtrnntov.telecom.com.au>
Derek Wildstar <email@example.com>
Aaron DaMommio <firstname.lastname@example.org>
Phil Pugliese <PPUGLIESE@pimacc.pima.edu>
Kenneth G. Hagler <email@example.com>
Michael Llaneza <firstname.lastname@example.org>
Hans Rancke-Madsen <email@example.com>
From: firstname.lastname@example.org DATE [Xboat 302, 303, 304, 305, 309, 310, 312]
Download date: DD MMM YY
Subject: TCS Design Strategy
[Nick’s original posting, 302]
I was reading E.K. Drexler's "Engines of Creation", learning a bit about nano-technology, when I came across this little gem. If it's been posted before, please forgive me...
Prof. Lenat at Stanford University developed an AI program. Lenat trained this system, EURISKO, to produce Trillion Credit Squadron designs and compare them in combat against each other. Drexler continues (p73, OUP edition):
"Lenat and EURISKO entered the 1981 national Traveller TCS tournament with a strange-looking fleet. The other contestants laughed at it, then lost to it. The Lenat/EURISKO fleet won every round, emerging as the national champion. As Lenat notes, `This win is made more significant by the fact that no-one connected with the program had ever played the game before the tournament, or seen it played, and there were no practice rounds.'
"In 1982 the competition sponsors changed the rules. Lenat and EURISKO entered a very different fleet. Other contestants again laughed at it, then lost. Lenat and EURISKO again won the national championship.
"In 1983 the competition sponsors told Lenat that if he entered and won again, the competition would be cancelled. Lenat bowed out."
The question is, did JTAS ever publish notes on the winning squadrons, and if so, what did they say? [Did they ever publish stats for the winning squadrons, as half-threatened, or were they too embarrassed?]
University Chemical Laboratory, Lensfield Road, Cambridge, CB2 1EW, England
Subject: TCS Design Strategy
[Mark’s comments, 303]
Lenat's design from 1981 was published in JTAS #10, pp. 38-9. He won at Origins that year. There is no reference to the computer design of the ships, just that it was the "most unique" fleet.
Not being a TCS player, I have no idea how "most unique" his design was - I'll post it here if there is any interest.
Subject: RE: TCS Design Strategy
[Django’s comments, 303]
Nick Munn <email@example.com> writes:
> I was reading E.K. Drexler's "Engines of Creation"...
It hasn't been posted on this incarnation of xboat anyway. I also came across this in the above book a few years ago and remember thinking "So THATS why TCS tournaments stopped so soon".
> The question is, did JTAS ever publish notes on the winning squadrons, and if so, what did they say?
No, JTAS never got around to publishing the winning TCS squadrons but they did have a few articles about the design of an efficient Trillion Credit Squadron.
I do have a listing of TL12 ship designs that are apparently those of a winning TCS fleet that I obtained from the HIWG. It's rather long so its probably better if you email me if you want copy.
Subject: Re: TCS Design Strategy (XBOAT#302)
[Derek’s comments, 303]
Nick Munn wrote:
> The question is, did JTAS ever publish notes on the winning squadrons, and if so, what did they say?
Not as far as I can recall (I have all issues of JTAS since #4, and I only recall one TCS design article, which does _NOT_ seem to refer to the Lenat/EURISKO fleets - it was probably written by a later winner).
Personally, I feel that by _not_ publishing the winning squadron designs GDW did TCS/High Guard players everywhere a disservice - it's quite possible that by studying the designs and the reasons they were so successful that we could have learned something about the TCS/High Guard system, and possibly been able to meet the program on it's own terms in future competitions (or at least, exploited the same loopholes in the design sequence that EURISKO did).
One wonders, though: if the same person (human) had done the winning designs for two years running, would GDW have barred that person from competing in subsequent years?
Would anyone happen to know what these designs looked like?
This also points up what I consider to be a major weakness of the High Guard combat system - there's very little "maneuver". As a High Guard admiral, your major decisions are what to put in the line versus the reserve (and when to move ships from one to the other), and to designate targets for your ship's weaponry. This emphasizes the importance of the ship design, of course.
An improvement (and in this regard, I _do_ feel that the current offerings from GDW _are_ an improvement) is to incorporate enough maneuver and other tactical options that a skilled commander in a merely average ship has a good chance of besting a merely average commander in a better ship.
BL/BR isn't perfect - there should be sufficient design options that if you _know_ what type of ships your opponent is going to have, you should be able to design a fleet that will counter it. (This would have been the best option for dealing with EURISKO - a smart player, or someone else with access to a lot of computer time - could design a counterfleet for the design that the program was most likey to come up with, or possibly EURISKO would have stated that there was no 'best' fleet).
"It's Science Fiction, if, presuming technical competence on the part of the writer, he genuinely believes it could happen." --- John W. Campell, Jr.
Subject: EURISKO's designs
[Aaron’s comments, 304]
I read an article about EURISKO in Scientific American or some such mag many years ago; it leaped out at me because of the discussion of TCS. As far as the program's strategy, all I remember was that the program was supposed to have designed fleets that were uniform -- all the ships were identical. This meant that the loss of any one ship caused no loss of features from the fleet, only a proportionate loss in firepower.
I've never played TCS, so can anyone comment as to whether this is a chestnut of TCS design or a useful design concept?
Subject: Re: TCS tourny winners (fwd)
[Mark’s comments, 304]
On Wed, 14 Jun 1995, Michael Llaneza wrote:
> I'd like to see these, however I'd suggest putting them on a Web site and putting the address on XML.
I'm not sure about posting to a Web site, since this is copyright material. If someone can get GDW approval, do it, I'll send a copy of the designs to anyone who wants to type them in.
By the way, the designs use L-hyd drop tanks, and I'm not sure these are legal under the later tournament rules - any feedback?
Subject: Re: TCS tourny winners (fwd)
[Phil’s comments, 304]
> By the way, the designs use L-hyd drop tanks, and I'm not sure these are legal under the later tournament rules - any feedback?
Well, if memory serves me right, 2 out the 3 battleship designs (Tigress & Kokkirrak) in the CT 'Fighting Ships' only had enough internal fuel space for J3, despite their J4 jumpdrives. The Plankwell *did* enough internal fuel for J4 but it also had a wimpy meson screen hobbling it.
I once wrote a BASIC prg, way back when, to help me design HG ships. I found that if big BB's only set aside the 30% of internal space needed for J3, they had ample room, at TL-15, for the best of everything. If 40% was set aside for J4 then something had to be sacri- ficed.
After playing around a little I concluded that the limit of one spinal mount per ship dictated that designs be produced that would yield the maximum number of spinal meson guns. I think some others here have posted to that effect previously.
In my universe I decided to change the rules so that meson guns could be defended against better than the official rules allowed.
Subject: Re: TCS Designs
> I've never played TCS, so can anyone comment as to whether this is a chestnut of TCS design or a useful design concept?
The TCS fleet published in JTAS #10 is not uniform - there are four types of major combatants and two types of fighters. However, there are 75 of one of the major combatants (11,100 tons), with 11 total of the three other major combatants (four at 12,000 tons, four at 19,980 tons, three at 19,600). The fighters are really more small battle riders (seven at 1,000 tons; three at 100 tons).
I'm not much of a TCS player, but it looks like the fleet relies on missiles (there are only four meson guns and three particle accelorators in the whole fleet).
Subject: RE: TCS Fleets
I found the designs below in a much larger HIWG file which included High Guard stats for just about every CT ship GDW used in their products. I remembered that file as being large but the TCS part is so small that I thought I'd post it anyway.
Winning TCS fleet - TL 12
Four Garter class:
TB-Garter TB-K1567F3-B41106-34009-1 MCr 17,584.104
Bearing C 1 EE 7 12,000 tons
Batteries C 1 EE 7 crew=170
Agility=4; Fuel=840; Cargo=4.3 low=170
Note: L-Hyd drop tanks add 6000 tons of fuel and mass,change the agility to 4, and cost MCr6.01.(TB-K1344F3) The ship is designed to manuever when carrying up to 72,000 tons of drop tanks and one Wasp fighter.
Four Cisor class:
BD-Cisor BD-K9525F3-E41100-340C5-0 MCr22,291.175
Bearing 1 11 1U 19,980 tons
Batteries 1 11 1U crew= ?
Agility=0; Fuel=999; Cargo=19.1 low=95
Note: L-Hyd drop tanks add 9,990 tons of fuel and mass, and cost MCr10. (BD- L9313F3) The ship is designed to manuever when carrying up to 29,970 tons of drop tanks.
Three Queller class:
BH-Queller BH-K1526F3-B41106-34Q02-1 MCr27,802.392
Bearing Z 1 NN1 N 19,600 tons
Batteries Z 1 NN1 N crew=263
Agility=0; Fuel=1,176; Cargo=10.72 low=232; marines=200
Note: L-hyd drop tanks add 9,800 tons of fuel and mass, and cost MCr9.81. (BH-L1314F3) The ship is designed to manuever when carrying up to 29,400 tons of drop tanks and two fighters (one Wasp and one Bee).
Seventy-five Eurisko class:
BA-Eurisko Ba-K952563-J41100-34003-0 MCr13,030.385
Bearing 1 11 V 11,100 tons
Batteries 1 11 V crew=131
Agility=2; Fuel=555; Cargo=8 low=0; marines=35
Note: L-hyd drop tanks add 5,550 tons of fuel and mass, change the agility to 1, and cost 5.56. (BA-K931363) The ship is designed to manuever when carrying up to 16,650 tons of drop tanks.
Seven Wasp class:
IL-Wasp Il-A90ZZF2-J00000-00009-0 MCr896.75
Bearing 1 1,000 tons
Batteries 1 Crew=19
Agility=6; Fuel=60; Cargo=0 low=0
Three Bee class:
FF-Bee FF-0906661-A30000-00001-0 MCr127.945
Bearing 1 2 99 tons
Batteries 1 2 crew=2
Agility=0; Fuel=5.94; Cargo=0
Subject: commentary on winning TCS ships
[Kenneth’s comments, 310]
I just had a chance to look over the winning TCS designs that Django posted Tuesday, and had some comments on them.
Firstly, I can see why people laughed at them.
Okay, they've all got lots-and-lots of fuel, but except for the Wasps they're so lacking in maneuverability that their opponents would have to try to miss them.
They also seem to have lots of relatively useless energy weapons. I suspect that may be to take advantage of the High Guard quirk about weapon hits knocking off one factor at a time if there's only one battery of a given weapon...
Some comments on the individual classes--
Garter class: I can't quite figure out what this class is for exactly. The 'TB' designation suggests it's a tanker, but it's pretty heavily armed (and fast) for that. And why bother with tankers for a one-shot tournament? Maybe it's supposed to confuse the enemy. :-)
Cisor class: Pathetic speed and no maneuverability. It's armed nicely, but because it's a sitting target it's opponents should be hitting it much more often than it can hit them.
Queller class: Same speed and maneuverability flaw. This has lots of incredibly wimpy batteries and a big PA mount. To bad all the little batteries don't much chance of hitting a target with respectable maneuverability. It's also _really_ well supplied with sandcasters, all to weak to do _anything_ unless the enemy's guns are just as wimpy.
Eurisko class: These are apparently supposed to be the mainstay of the force, since there's so many of them. At least this class has _some_ maneuverability. It's extremely well armoured, so at least it can jump to safety once its weapons and m-drive are gone. Not that it will have enough fuel without its drop tanks. Once again, it has lots of really weak batteries, which combined with its poor agility mean it will be taking many many more hits than it delivers. Finally, its lack of a fiberoptic computer makes it very vulnerable to meson gun hits.
Wasp class: The USP for this gives m-drive and plant as 'Z.' From agility and fuel, I'd guess this should be 6. This is the largest 'fighter' I've ever seen--on par with an Alliance Navy Fast Patrol Boat. I actually like this one--probably because it's basically a low-tech version of something I came up with for Neubayern in Steve Higginbotham's TCS game. My PF design never saw combat, though.
Bee class: This is more like the heavy fighters developed towards the end of Steve's game. Not a very good example of the concept, though. Again, it's unmaneuverable, and armour on a fighter was found to be useless (and wasteful) in practice. Interestingly, it appears to have two seperate missile launchers in one turret _firing as seperate batteries._ This is a very nice idea--I wish I'd thought of it. I don't see why the sandcaster wasn't another missile, though...
Finally, this whole squadron has a flaw you could fly a Tigress through: not one ship can jump without drop tanks. So, all an enemy has to do is fire a salvo at each ship, blowing off their drop tanks, and then calmly withdraw to a safe distance. 30 days later the whole force runs out of power plant fuel and has to surrender. :-)
Of course, that doesn't matter in a one-shot tournament, but I can see how it would annoy any other entrants who might have designed more practical ships.
Now, I may be missing something here, since they did win. But, how? Their design philosophy seems totally at odds with what worked in an actual TCS game. I'd love to see how this force would fare against a TL12 force from Steve's game...
Kenneth G. |
firstname.lastname@example.org My insurance company |
Finger me for PGP 2.6ui public key is Beretta U.S.A. |
Subject: TCS tourney ships
[Michael’s comments, 312]
In looking at these designs from a Battle Rider perspective instead of a High Guard perspective they look rather different.
I am going to be converting these ships to BR terms, and in looking them over I've got one observation to make right off:
Missiles kill things.
Direct fire weapons of any kind in BR just can't punch through the defenses of a comparable ship. Lining two identical ships of the line (CLs or bigger) up and firing them at each other yields a long series of No Effect since its as easy to make tough defenses as it is to make Big Guns.
Missiles on the other hand, just plain do damage if they hit. To anything.
These ships have lots of missiles (I haven't come up with the exact numbers yet) and should do very well against any opposing fleet.
The low manoeuvrability of these ships is also less of a handicap after they get translated since the idea is to go active to guide missiles only when the missiles are about to hit, thereby giving your opponent as little time to shoot at you as possible.
Finally, I'd like a wee bit of advice on converting High Guard ships to BR:
1) As I recall the Agility of an HG ship is roughly its manoeuvre G’s in combat. Is that correct?
2) Fusion and plasma guns no longer function as anti-spacecraft weapons since the ranges have gone WAY up. So, Should they be replaced with laser turrets, point-defence lasers to shoot down all those missiles, particle accelerators with low range and high damage, or (probably the most popular) should fusion and plasma weapons be added to BR (this way you don't have to go near TNE/FF&S to play BR)?
I'll keep track of all replies to this and eventually post a summary. My intention is to create a _loose_ set of _guidelines_ to enable the easy conversion of High Guard ships to Battle Rider. I am *not* interested in writing a bulletproof set of rules for doing so, I don't have the time and BR doesn't need that much detail.
Thanks for your attention, and sorry for corrupting XML with a New Era product.
Michael Carter Llaneza
Conceptual Design Services The Worse it gets,
Pi Kappa Phi The more I get used to it.
"I am the NRA" Duty Now For The Future
Veteran of the Bermuda Triangle Expeditionary Force 1990-1951
Subject: Winning TCS design
[Han’s comments, 312]
Kenneth G. Hagler writes:
> Now, I may be missing something here, since they did win. But, how?
I suspect that all those crappy extra weapons are not really weapons, but armour. Under the unmodified TCS rules the defender decides just what system in a category that gets taken out by a hit.
> Their design philosophy seems totally at odds with what worked in an actual TCS game. I'd love to see how this force would fare against a TL12 force from Steve's game...
I think they'd lose, but then, Steve made a lot of changes to the original TCS rules. He'd allow anything that made sense. Just imagine the difference that'd make! ;-)
University of Copenhagen
"A subsector official pompously states that the subsector armed forces have four Kinunir class ships in service, each with enough troop strength to put down any military operations that threathen the peace of the Imperium." - Adventure 1, The Kinunir