My first rules botch- Stratego

My mother has a battered set of Stratego, an older version released sometime in the late 80s or early 90s. A blue bomb piece was missing, but we usually compensated by making red remove a bomb as well. I was taught by my mom, and frequently played with my cousin Sharon while she babysat me as a young kid.

Stratego seeks to emulate Napoleonic Wars of attrition, featuring a constrained map where pieces move one square at a time and the main approach to your enemy is three two-square wide channels with impassable lakes or the board edge to the left and right.

The victory conditions in Stratego are simple- kill every piece your enemy has that can move (Flags and Bombs are immobile), or capture the enemy's flag.

Stratego is a meat grinder. The strength of pieces is known only to their player- and all pieces are rated on a 10-point scale for strength. In older versions, smaller numbers were stronger, much like AC in 2ed AD&D (Which I'll use here). You can only discover the strength of opposing forces by attacking- a risky venture as discovering that you just walked your 7 into your opponent's 2 means losing the 7. Ties result in both pieces dying in a Pyrrhic Victory.

My family generally played with the following understanding:

• Scouts (9) can move and attack in the same turn
• Bombs destroy any attacking piece that is not a Miner (8). When attacked by a Miner the Bomb is removed. Bombs are not removed upon killing attacking pieces.
• Victorious attacking pieces advance into the square of the defeated piece
• Victorious defending pieces do not advance into the square of a defeated attacking piece
• Moving a non-scout piece to the far end of the board (similar to a pawn advancing in chess) allows you to rescue a formerly captured piece, which is placed on your starting side of the field. This may be done 3x per game, but each individual piece may only rescue once.

But, Sharon and I didn't read the rules closely enough, because there's this cool piece called the Spy.

At the top of the Stratego foodchain, you have the Marshall (1). The Marshall is a force of nature, capable of cutting down swathes of troops, only challenged by Bombs and the enemy Marshall in the absence of the Spy. 

The Spy is squishy, and dies to any attack, while also dying when attacking other pieces. But, if the Spy Attacks the Marshall first, the Marshall dies. At that point the Spy can just sit around body blocking, preventing Scouts from blazing across the battlefield and taking unprotected flags, or something I guess.

But Sharon and I didn't understand the rules that way. We'd unknowingly played with a much deadlier set of rules:

If the Spy Attacks, it wins (except against bombs). If the Spy is attacked, it dies.

This made a game that was already a stark look at attrition and added a game of monstrous glass cannon chicken into the mix as well. The spy could annihilate and cut through an opposing enemy army, but any scout could quickly end its reign. Attacking the wrong bait could cause the Spy to advance into the defeated piece's space and fall victim to a piece lying in wait.

Previously, the Spy was best used hidden, unobtrusively placed to sneak out and get the drop on the Marshall, a contingency to the piece in the game meant to be a force of destruction.

With our rule, what was an elegant and subtle duel became a frantic dance to find an opening that would allow for taking out one of the deadly duo, without losing half of your own in the process.

Now that I'm significantly older, I've played Stratego with both sets of rules. As far as screwing up rules go, the unintentional change we made to the Spy wasn't game-breaking, unlike the scores people who never ran auctions for unpurchased Monopoly properties (let alone the insanity of allowing free parking to accumulate a pool of money).

I'd say that this rules botch is a viable rules variant. It can overall speed the time a game of Stratego takes because of the rapid death experienced by armies as the deadly duo comes through town, and it simplifies the Spy's role by making its interactions with other pieces more universal. Highly suggested when playing with young kids looking for a more action-packed board game.

GURPS Fail: Garrotes and Targeted Attack

Disclaimer: I'm aware that Targeted Attack is intended for ranged weapons- specifically firearms- by RAW.

I've been elbows deep in a character concept that might actually see the light of day, but in doing so I started looking at using a Garrote.

Garrotes are awesome because if you can eat the -5 penalty to hit the neck (and they are quite kind to suggest using an all-out attack for +4, meaning an overall -1 skill penalty), you have the enemy trying to escape at ST-5, or Judo/Wrestling at -3. Contests to strangle give you a +3 bonus to ST.

The ST bonus for you and the penalty for the enemy are fairly significant, especially since you contest with Garrote against the attempt to escape. It allows a low strength individual a fairly nice way of taking out higher strength enemies.


Now, weirdness ensues when you use Targeted Attack Neck (Garrote) to reduce the hit penalty to -2 (from -5).

Without Targeted Attack, assuming Garrote Skill 12:
12 (skill) +4 (all-out attack) -5 (neck) = 11 modified skill

With 4 points in Targeted Attack, assuming Garrote Skill 12:
12 (skill) + 4 (all-out attack) -2 (neck) = 14 modified skill.

With 4 points in the Targeted Attack Technique, the skill level of Garrote effectively increases by 3.


I think my big issue here is that Garrote is not a skill that works without targeting the neck. You don't attempt to Garrote limbs, you don't wrap rope around someone's torso, a Garrote goes around the enemy's neck, period. There seems to be little point enforcing a location penalty when you have no option but to target that particular location.

My Fix:
Garrote is now a DX/Average skill. Attacker rolls vs Garrote to get it around the victim's neck. No attack modifiers (All-out, Telegraphic, etc) are valid.

Benefits: Fewer modifiers to deal with, overall same skill levels, no possibility of improving it with a technique.

GURPS Fail: Digital Possession

I've been thinking about an AI concept lately- the AI is generally uploaded into a clumsy robot shell, but it can leave the shell and possess other machines to control them. Reading up on this lead to some unsatisfactory results.

From Basic Set p76, we have the Digital limitation to Possession:

Digital: This limitation is only available to Digital Minds (p. 48). You take
over computers, not living bodies. The
target system must be connected to
your current host computer via a network, and you must have complete
access to it – voluntary or otherwise
(see Computer Hacking, p. 184). The target computer’s hardware must be complex enough to run your computer program; in general, its Complexity must
be at least half your IQ (round up). You
can also take over a computer using a
copy of yourself while leaving the original intact! However, unless you have the
Digital version of Duplication (p. 50),
any system you take over this way
becomes an independent NPC that
thinks it is you. This can be good or bad
– the duplicate could become any type
of Associated NPC (see p. 31). -40%.

 A quick check of the campaigns book shows that 2004-era desktop systems are Complexity 4 (Basic Set p472), although super computers and the like would obviously be more complex.

Now, I understand the reasoning that a desktop PC couldn't handle running an advanced AI- but that's completely ignoring how fantastical digital possession would be in the first place. Such technology capable of completely overriding another system and taking it over would surely be capable of dealing with the Complexity issue.

The other big design fail here is that Possession seems strictly inferior to Mind Control with the Cybernetic Enhancement/Limitation:

Cybernetic: You can affect entities
with the Digital Mind trait (p. 48),
including all ordinary computers.
Your IQ roll has a penalty equal to the
system’s Complexity. A nonsentient
system does not resist; just roll vs. IQ -
Complexity to succeed. +50%.

So let's compare. Digital Possession with Full Memory Access, and Telecontrol 1 is going to run you 120 points. You're only able to possess machines with complexity equal to half of your own IQ- so in a TL8 setting your IQ is maxed at 8 if you want to possess a rudimentary desktop PC. Otherwise you're stuck possessing far more complex machines, all of which are going to be incredibly rare. You have to win an IQ vs. Will contest to succeed, even against non-sentient computers.

Mind Control with Cybernetic Only will run you 25 points. An IQ 16 AI with it will take over a desktop computer ~74% of the time with an unopposed IQ-Complexity check, and then be able to tell it to dump data logs, open the security gate, and delete the logs that the intrusion by the AI ever happened.

Even if you need a version of Telecontrol (wifi) to make Mind Control work with machines at a distance, there's no way in hell that's going to overcome the 95 point gap between the advantages.

While I'm ranting about GURPS:
Telecommunication (Internet/Wifi) never made it into the Basic Set or Powers? Really?

Danger Sense: Think more Intuitively

So Christopher Rice and Douglas Cole both posted about Danger Sense. They point out that it's passive, easily forgotten about, or becomes stifling for the GM to work around (which is the same reason Mind Reading is banned in all of my games).

But I'm not a big fan of a danger sense point system similar to luck, serendipity, and all of the other metagamey stuff you can by to help plot protect your characters.

No, I think you need to think more intuitively.

As in, Intuition.

Intuition allows a player to make an educated guess/hunch on a topic that doesn't build off of existing knowledge. The GM makes a secret IQ roll with a modifier equal to the number of "good" choices minus the number of "bad" choices. For example, if the PC is wondering which car in a parking lot full of seven cars is the one stashed with drugs, it would be a -6 to the check (1 right car, 6 wrong choices, 7 total).

Danger Sense can be much the same, where players have to actively decide to use the ability. Sure, this weakens the ability to an extent as a means of protecting a character played by a players who might not be very cautious, but Danger Sense also doesn't inherit penalties to perception rolls like Intuition does with IQ rolls. A character with high perception will reliably be able to tell that something is just... off. Any other character wouldn't even have the option of trying such a thing without having a relevant skill, much like characters can't attempt intuition without the advantage itself.

Now, for limiting Danger Sense, I have two main suggestions:

• Only allow one Danger Sense roll per situation/scene. An entire combat is likely a scene, sweeping a warehouse is likely a scene, a meeting is probably all a single scene. Now if a gunship helicopter suddenly arrives during any of those scenarios, that might warrant being able to check again to see if the character gets a sense that maybe they should get the hell out of dodge.
• Limit perception for use with Danger Sense to ~14. Characters can have higher perception, but for Danger Sense checks, anything 15 and above fails. (Although I'd be tempted to limit to 12, which is a 74.1% chance of success, where 14 is a 90.7% chance of success.) Alternatively, you could assess penalties and bonuses on situation, but that's probably up to personal preference.

Defeating Security: Some EXCELLENT videos

I recently did a post on defeating doors, and I gave keypads and other electronic mechanisms only passing mention.

I stumbled on a Vimeo channel that really shows how a character such as Tobias (yes, all the character sheet images are gone, it's a post from almost two years ago) could believably beat any number of security devices given access, time, and the correct information.

Safe Cracking (Vimeo videos can't be embedded on blogger evidently, since Google wants us to be linking youtube, which they own.)

In the first video, the guy cracks open a safe using a screwdriver and a paperclip.

Electronic bypass of a safe lock

Now, what the videos don't show is how much time it took to initially look at the safes and recognize the security flaws in them- but instruction manuals abound on the Internet, potentially reducing the time needed to overcome such flaws. I'm sure this also gets much easier with practice as well.

Possession of the manual makes this safe a joke (Bonus use of a straw for circumventing security)

What really strikes me here is that a lot of consumer products being marketed as being very secure (after all, guns are serious business) are flying under the radar as far as quality goes- meaning that people shooting for less security are just as exposed.

Resetting an access code

(Side Note: 4 button and 3 button combination locks provide 24 and 6 different button combinations respectively, a laughably small amount that's easily brute forced. At minimum, 5 buttons gives 120 different combinations, which is at least another five minutes or so entering different combinations.)

Wafer locks are bad and McDonalds plastic forks/knives helps crack a safe

Personally, I feel Lockpicking and Electronics Operation (Security) don't reasonably cover the bases in a way I find satisfying. Picking a door lock is one piece of physical security circumvention that Electronics Operation doesn't fully cover either. I feel Electronics Operation strongly implies that such circumvention is done on a software level- abusing programming to get it done, rather than relying on hardware weaknesses.

Forced Entry isn't a viable choice either, seeing as how it's about smashing and grabbing.

What is obvious is that working without tools for physical security circumvention is nigh impossible- easily -8 or -9 to skill, with even improvised tools reducing that a great deal (after all, a paperclip is still an improvised tool). DX is vastly more important once you know how to circumvent a device, but IQ is still definitely needed to understand the problem at hand.

Getting Past Doors: Modern Options

Dungeon Fantastic has had a preoccupation with doors lately, and I figured I'd throw out some ways of defeating doors in the TL 6+ world. Most of these apply to generic locked doors that are largely intended to keep the honest person honest- not your high-security fare. Truth be told, almost all of these focus on the lock, not the door.

Bump Keys

A bump key is a key that has all of the teeth at the shortest possible length- 0. Most locksmiths or hardware shop folks will look at you really funny if you request one of these. Depending on jurisdiction, Bump Keys could be illegal, probable cause for a search, or an item that can be confiscated by police.

Using a bump key is simple- insert the key one notch out, apply sideways pressure to the key in the direction the lock opens, and bump the key into the lock making sure the motion engages the teeth into the lock. 

Bumping is fast and relatively quiet. Bumping doesn't leave many signs of forced entry, especially done with a light touch. Less careful bumping can leave telltale marks on a lock, although your normal beat cop might not pick up on them- I figure most detectives or locksmiths would.

Bumping is limited largely by needing the correct key base to put into the lock. Countermeasures almost all rely entirely on the quality of the lock involved, and generally involve needing specific extra precautions be taken.

For use of a proper bump key in GURPS I'd give a hefty equipment bonus to lockpicking skill- on the order of +5 to +7. This assumes that the bump key is suitable for the door in question, and that the lock doesn't have any countermeasures against bumping.

Credit Card

My workplace is a shared complex with multiple businesses, many of them with dozens of employees. For the longest time, there was a single men's and a single women's room, and the men's room had one stall.

Lines were frequent. Trips to other buildings nearby to use their facilities were common.

Luckily, an old tenant gave me a tip about the location of a hidden bathroom, behind the mysterious door K.

Door K has a very simple lock, no deadbolt, and the lock doesn't have any covering or shield preventing you from sliding a credit card between the latch and the door frame.

Using a credit card or other shim on a door is about as quiet as bumping a door. Success really depends on having good access to the latch, and an appropriate shim (my friend has used a spatula when I wasn't around to bump key open her room for the second time she locked herself out).

The most common counter measure to this method of cracking a door is a plate that hides the latch from outside interference (See example to the right).

Against doors that are vulnerable to being opened with a Credit Card or Shim, I'd give a bonus of +2 to +4 to Lockpicking skill.

Keypads

The company that has started to glob up the majority of the vacant rooms in the complex I work at installed keypad locks on all of their doors lately. Sadly for them, Keypads are weaker to social engineering and bad operational security than a physical key.

With time and the inclination, opening an electronic keypad and creating a short is definitely within the realm of possibility- although it will be obvious and time consuming, and higher quality locks won't open with loss of power/a short. Electronics Operation (Security) and/or Electronics Repair (Security) are both really useful here.

Alternatively, getting the key code from someone is far simpler in most cases. All it takes is overhearing one person telling someone else the code during orientation, or seeing the right email.

Some locks with keypads don't actually proof against bumping- they only offer an alternative path to opening the lock without removing the original key mechanism already present. Often times manufacturer documentation can provide master passwords and the like that can circumvent security measures.

The manual listed the default password- which had never been changed,
allowing me to change the sign to whatever I wanted

Piggybacking

Want to get into an apartment building? Loiter around the front until someone comes in or out, and slip in while the door is still open. If needed, claim that you forgot your key, that your girlfriend is late, or that you forgot your wallet. Piggybacking is easy, requires no tools, and the only risk is that whoever opened the door for you doesn't get suspicious.

College campuses are by far the best location for piggybacking through locked doors.

POWER

Numerous power tools have uses in defeating locks/doors. All of these are loud, with some being EXTREMELY FUCKING LOUD.

• Many locks are vulnerable to being drilled
• A circular saw appropriate for cutting metal can cut through latches/hinges. This is slow, extremely loud, and potentially dangerous for the saw operator.
• Thermal torches can cut through latches, or can be used to make holes in doors. 
• Jacks can be used much like rams, except instead of swinging a log you allow pneumatic or hydraulic power do the work for you

Firearms

Shooting a lock can be a really hit or miss proposition. Most breaching done these days is done with a shotgun loaded with a breaching round. The muzzle of the shotgun is placed between the lock and the latch, pointing downwards at a 45 degree angle to minimize the chance of the round hitting unintended targets, and then the round is fired.

All told, if you're going to use a firearm, go with shotgun slugs or a rifle, and be prepared for the possibility of shrapnel.

This is backed up by the stats given in High Tech for Locks. (HT 203). A standard lock has DR 6 and HP3, plus unliving injury tolerance. A 9mm pistol will take 2-3 shots to break such a lock- rolling max damage each time.

In Closing:

Modern breaking and entering definitely has more subtle options available for bypassing physical security, especially once magic is no longer an option. Most of the very loud options are gratifyingly loud to boot, which can really amuse players.

GURPS: Vehicle Hit Locations for Cars Analysis

Precis- I explore some of the wonky oddities surrounding Vehicle Hit Locations

First, this post on explosives is excellent (didn't know Roger had a blog until an hour ago).

I'm going to take a look at vehicle hit locations. This is something that has come up before in Agency 17.

Compared to helicopters (control surfaces), planes (ditto), spacecraft (who wants to inhale vacuum?), and boats (water), cars don't care too much about damage to the exterior. Shooting up the side panels, passenger compartment, or the trunk of a car does jack shit towards disabling the vehicle. The way I see it, a car only has a few real vulnerabilities:

• Engine
• Wheels
• Fuel Tank
• Driver
• Drive train (really starting to get pedantic here)

So let's look at the actual table, piece by piece. I'm going to focus specifically on cars. That means I'll be ignoring things like superstructure, wings, exposed riders, etc.

Body: Powered vehicles that take a major wound must check vs, HT or suffer halved move.

Guinea Pig:
The TL6 Roadster (B464) has ST/HP 42, HT 9f (flammable), SM +2 and DR 4, and counts as unliving.

A piercing attack needs to deal 21 points of injury to cause a check for the major wound. That would be 63 points for a piercing attack (unliving reduces incoming damage by a 1/3rd), 42 points for a large piercing attack (1/2 reduction via unliving), or 21 points of damage for huge piercing or impaling attack.

A quick look at TL5/TL6 rifles in High Tech shows a few that could manage to knock out a car in a single body shot, accounting for -4 damage due to DR:

The TL5 Greener Elephant Rifle (although admittedly extremely rare in Prohibition Chicago) can  pass the test with 6d+2 pi++ damage, a maximum of 34 (after applying for DR 4), 8 points of damage shy of the threshold large piercing attacks have to hit to cause a major wound. A quick look at http://anydice.com/ shows that 6d+2 will deal 21+ damage 36.31% of the time.

The TL6 H&H Royal Double-Express, .600 NE (also extremely rare in Prohibition Chicago) also easily does the job with 5dx2 pi++ damage, doing at least 21 damage roughly 3.24% of the time- before figuring in the x2 portion of the damage text which makes it a much more likely 77.65% chance.

The Steyr-Solothurn S18-1000 is more or less a souped up H&H Royal Express with 6dx2(2) pi++, dealing 22 damage 99% of the time.

As for machine guns:

The Maxim 1-pdr is an interesting case due to the armor divisor of (0.5), giving our poor roadster effective DR 8. Still, with 5dx2(0.5)pi++, it can manage to cause a roadster a major wound 39.97% of the time! This isn't even accounting for the followup explosions (which will wreck any occupants extremely fast).

The Browning M2HB makes the cut, dealing 42 points of large piercing damage roughly 50% of the time. The KPZ DShK-38 has identical damage numbers.

A 12-gauge 2.75" firing slugs (4d+4 pi++) can also hit the 21 damage threshold 2.7% of the time.

Verdict:

The body location mechanics make a good deal of sense- only a small handful of TL5-6 rifles can pump out enough damage in a single shot to potentially knock a vehicle to half move, and two of the three machine guns available are definitely WW2 offerings. Shotgun slugs can also do it, but with abysmally small chances.

Notably, some favorite TL8 anti-material rifles such as the Barret M82A1 don't pack enough wallop to disable a TL6 roadster with a single body shot.

 Glass windows, Large and Small

When you hit a glass window, you make an Occupant Hit check and if that succeeds your attack hits someone inside. Depending on if the window was up or down, they get half vehicle DR or no DR respectively on the attack.

First, there isn't much description as to the size different between a Large Window (a -3 penalty to hit, reduced to -1 for the roadster's SM) and a Small Window (-7 penalty to hit, -5 for the roadster). If a full windshield is a -1 to hit, would a normal car window really be at -5? I mean, the size difference between all the vitals for a human and the skull hit location is much smaller than a full-sized windshield to a car door window, right? It's a maddening bit of ambiguity.

Using our Roadster example, a single occupant would be hit on a 9 or less (37.5%), while two occupants have a 50/50 shot of either of them taking a hit on the 10 or less. Now, this is the straight attack, not the 5 points becoming 1d of cutting damage under Occupants and Vehicle Damage- your 7d+1 Springfield rifle shot plows into someone.

p.470 also includes the following on windows:

An attacker can sometimes target a
vehicle’s occupants directly. This is
only possible if the vehicle has an
exposed rider (E), glass windows (G or
g), or an open cabin (O)...  A rider has no cover; someone in
an open or glass-windowed vehicle
has partial cover (legs, groin, and half
the torso). There is an extra -1 to shoot
into or out of a window unless the
occupant is actually leaning out.

Verdict: Why shoot a window when you can shoot the guy behind it at a -1 penalty? Shooting through the vehicle DR and hitting your main target is a far better bet than relying on Occupant Damage.

Vital Area: 

Vital Area: A powered vehicle (anything with a ST attribute) has vital
areas: engines, fuel tanks, etc. The
wounding modifier for a tight-beam
burning attack is x2; that for an impaling or any piercing attack is x3!

This is where things get really wonky in a way that really rubs me the wrong way. The quote above seems to strongly indicate that Vital Areas do not benefit from Injury Tolerance: Unliving, a trait that most vehicles have. Hilarity ensues.

The M1903 Springfield rifle does 7d+1 pi damage. According to the entry for Vital Areas, it would receive a wounding modifier of x3. 14 damage after DR will hit the 42HP needed to reduce a vehicle to 0 HP immediately- 18 damage before DR. The M1903 will do 18 damage roughly 80.83% of the time with a single shot.

A 12-gauge 2.75" shotgun firing a rifled slug (4d6+4 pi++) will deal 18 damage before DR 15.9% of the time, also reducing the roadster to 0 HP in a single shot. 

I should also mention that dealing 42 points of damage to a vehicle in a single hit will cause 8d of cutting damage inside the passenger compartment due to the 5 points of damage -> 1d of cutting conversion as per vehicle damage and occupants.

Verdict: This is broken.

Wheels:

Damage over HP/(2*Number of Wheels) cripples a wheel. For the TL6 roadster, that's 6 points of damage. Effects of crippling a wheel are the same as a character with an equal number of legs losing one leg.

So uh, time to look up the rules for adding legs or losing legs in Characters? Because looking up extra legs to determine the properties of a vehicle losing wheels is intuitive and user friendly?

A quick look at B.54-55 shows that:

• For 3-4 wheels, crippling 1 reduces move by half.
• For 5-6, each missing wheel reduces move by 20% until only 3 or left (40% move at that point), and then any gone after that stops movement completely
• For 7+, each lost leg reduces move by 10% until 3 are left (40% move at that point), and then losing another stops all movement completely.

Additionally, any damage to a wheel forces an HT check. If it fails, it's immediately crippled.

Hitting the wheel of a SM +2 roadster is just a -2 penalty.

Verdict: Wheels are a primo target for ending car chases, especially since 4-wheel cars are the norm for most modern settings.

Insurance Claims:

Of the hit locations, the body hit location does the best job of mirroring reality- very few weapons available at the same tech level as the sacrificial lamb can reliably knock it out in a single hit, which I feel is very true to life.

On the other hand, the "Vital Areas" hit location is an abomination that seems to completely ignore  that an engine would have Injury Tolerance: Unliving, just like the rest of the car.

Wheels are an interesting case, and I feel that having one become crippled should force a control roll or  give control rating penalties and that the move penalty should be reduced to 25% until two wheels are crippled. 

Windows are another strange case because you can explicitly aim for a window and take your chances that an occupant inside gets hit, or you can just accept a -1 (unless you want to hang out your window like a chump to fire your gun) to try and hit someone inside, who gets partial DR for half their body.

Fix 'er up (my thoughts on potential fixes/solutions and balancing):

Body: Leave as is.

Wheels: The first crippled wheel on a four-wheeled vehicle reduces max speed by 25%. A crippled wheel requires a control roll to prevent a mishap/accident (I highly suggest applying speed penalties using everyone's favorite table to the driving check). A second crippled wheel reduces max speed by 50% and requires another control roll. 

A penalty to handling would probably also be an excellent idea as an alternative to maximum speed reduction or in addition to a maximum speed penalty. 

Windows: Attacking a window directly is a -3 penalty while attacking an occupant directly through a window is a -5 penalty. Choosing to attack an occupant directly dictates who is hit on a success, choosing to attack the window directly means rolling for Occupant Damage on a hit. If that subsequent roll is also a hit, a random occupant is hit by the original attack. If the Occupant Damage check is not a hit, the vehicle sustains damage from the attack, but do not make additional Occupant Damage Checks for the same attack.

If random hit location is used for a direct hit on an occupant, the legs, groin, feet and half of the torso (1-3 on a 1d roll) are protected by vehicle DR. A closed window provides half Vehicle DR.

Vital Areas:

Reduce the wounding modifiers for piercing attacks to a vital area:

• Huge Piercing: 1.5x
• Large Piercing: 1.25x
• Piercing: 1x
• Small Piercing: .5x

Impaling attacks get a wounding modifier of x1, because who is going to stab an engine block with a spear?

Attacking the engine has a -3 penalty and may only be done from the front or sides.

Attacking the fuel tanks has a -6 penalty (Exception: Externally mounted tanks, -3)  and may not be done from the front on most civilian vehicles. Damage over Vehicle Max HP/5 cripples the tank, causing the Vehicle's range to drop by 75% of its max.

Overall Vehicle HP and Damage:

Gurps currently doesn't have any stated rules (that I've been able to find with a more than casual less than decent search of various source books) on how to handle when a vehicle is disabled from generic damage. I'm willing to bet I've missed something, somewhere, but until I learn about something, I'd probably go with the following:

Once a vehicle reaches 0 or less HP, it needs to check against HT immediately and every turn afterwards or immediately come to a stop.

Vehicle Occupant Damage:

I would make a few changes:

• Called shots to the wheels and fuel tank do not check for occupant damage
• Shots that hit the engine from the side of the vehicle do not check for occupant damage
• Instead of randomly choosing who is hit, a GM may at their discretion cause Vehicle Occupant Damage to affect whoever in the vehicle is closest to where the damage to the vehicle occurs

Thanks:

I'd like to thank Douglas Cole for taking the time to read a draft of the post and getting in touch with David Pulver, who explained what lead to the vehicle hit table during development of 4th ed:

However, in the original draft of 4e, all hit points scaled with square root (like ST) rather than cube root. This had a significant
effect on vehicle hit points. A car, for example, would have had around twice as many hit points as it ended up with and a tank would have
about 3 times as many HP...  

The late change to cube root HP in the final draft, made after I had left the project and was thus unable
to comment, unfortunately affected some rules in subtle ways. The high multiplier for piercing damage was originally intended as a way of allowing
"golden BB" type effects, but with only half to a third as many HP in the final version produces some unfortunate overkill results.

Single Dice Resolution Systems Suck

A +1 boost in  a d20 system is a teeny 5% boost to your chances of success. That is because a single dice produces flat results.

Flatter than an A-Cup Angst Anime Character

Every possible roll on a d20 is equally likely. Systems attempt to account for this by moving the goalposts of what you need to roll over to succeed, accounting for bonuses to calculate the chance of success based off of the bonus you have.

After all, a +10 bonus guarantees you will hit a goal of 20 half the time.

3d6 resolution is different.

Curves like the girl who doesn't reciprocate my feelings

See that bulge in the middle? The tendency for results near the middle of the range of possible outcomes to be more frequent? That right there is awesome. It's still random, yes, but now skill bonuses matter- and become more interesting.

An effective skill level of 10 in GURPS represents a 50.0% chance of success (p.171 is amazing). A +1 raises that to 62.5% chance of success- an increase of 12.5%. That's two and a half times more bang for your +1 buck vs. a d20 system.

Now, the further you move away from the middle, the less the increases in skill matter. It's not until skill level 14 at the top of the scale that a +1 bonus in GURPS is less powerful than the same +1 bonus in a d20 system game.

The change in percents as your reach extremes is great- if you're extremely low-skill for some reason, jumping a level up is a tangible boost to success- while bonuses at the top have less and less effect, eventually becoming moot (17 and 18 always fail). In GURPS this frees you up for things like harder combat maneuvers, techniques, dangerous shit... in d20 it uh... why would you ever stop accepting bonuses to do rad stuff? You're losing 5% each time, which is trivial, sure, but you're still shooting to get above the goalpost, where in GURPS you can stop worrying once your skill hits 16 most of the time.

The goalposts you're aiming to roll under in GURPS vs rolling over in d20 aren't nearly as arbitrary. There's not a great deal of math involved figuring them out- just looking at the bell curve distribution anchors the resolution mechanic to a norm- you can expect a lot of 8-12's being rolled.

Special shout-out to 11d10-10 as producing a very beautiful 1-100 bellcurve distribution of roll values.

Allowing allies to take advantage of Feint

So, the combat in my game is starting to get slightly stale for my tastes, and I've always been slightly irked by how much better deceptive attack seems compared to a feint- the improved turn economy of deceptive attack is highly desirable.

A feint, done right, draws your opponent into a state of weakness so that you can then smash in their face, taking advantage of it.

My thinking is, a good enough feint might open up a hole in an enemy's defense so glaring that even allies can take advantage.

Proposal:
Any feint that imposes a penalty of -3 or more to a defense roll can be taken advantage of by allies, not just the combatant who originally used feint.

Parry Missile Weapons as a Technique

For the new Demon Hunters campaign I'll be running, a player with Enhanced Time Sense has expressed interest in being able to parry musket shots with a rapier, using parry Missile Weapons (PMW).

Considering he has dodge twelve, getting an equivalent parry score verses a musket would require Parry Missile Weapons 28 (28/2-2=12). Remember, parrying a bullet is a -5 penalty to parry, and parry is Skill/2+3 for Skill/2-2 when trying to parry a bullet.

Getting a skill to 28 is prohibitively expensive, on the order of 76 points for a DX 10 character, a price that barely drops with increased DX or levels of Enhanced Parry/Combat Reflexes.

So I propose making PMW a technique:

Parry Missile Weapons
                                        Hard
      Default: Cannot be used at unimproved default,  Unarmed or Melee Weapon Skill -5
      Prerequisites: Unarmed or Melee Weapon Skill, cannot exceed prerequisite skill +4.
This Technique allows a character to parry thrown or missile weapons with a ready melee weapon. If you are wearing wristbands or gloves with DR 2+ or have at least this much natural DR, you can also parry with your hands. Your Parry Score is (Technique Skill/2+3), rounded down- based on this technique, not the underlying combat skill that this defaults to.

Modifiers: +4 vs large thrown weapons (axes, spears), +2 to parry small thrown weapons (knives, shuriken), no modifier to parry arrows, -2 to parry smaller low-tech missiles (crossbow bolts, sling stones, blowpipe dates). Enhanced Time Sense and other appropriate advantages (e.g., Precognition) allow you to parry bullets at -5.

With this technique, a fighter with ETS can pay ten points to be able to parry bullets at a -3 to parry. Those without ETS will still enjoy much better parries against large thrown weapons (parry +4), arrows (parry +2), and smaller low-tech missiles (no modifier to parry).

For many characters with decent statistics, dodge will likely remain a better defensive option, and this technique really only benefits characters with high skill with their weapon to begin with- although stacking it with Combat Reflexes and Enhanced Parry certainly makes it even more powerful.