I agree with Magmai that if it took a realistic amount of time, then it would suck big-time. At the same time, making a horse into a four-legged motorcycle (Zelda, Dynasty Warriors) doesn''t capture the spirit of having and riding a horse.

Maybe all horses could be ridable to some degree or another, but the controls would be a little loose at first, and tighten up as you rode it.

Also, it''s absolutely critical that once you form a good bond with a horse, you can stand on a rooftop and whistle, and your horse will move to stand underneath you so that you can jump to its back. Sweet.

I see a trend here to confuse lengthy training with depth of control and relationship.

The goal is depth of control and relationship. Nowhere is it stresed that lengthy training is required.

Lengthy training can be mitigated to some extent by buying or selecting or stealing a horse that already has been well trained. But going this route does not mitigate the necessity of improvement through bonding.

Another thing that isn't being realized here is that training happens through riding - somewhat automatically. Ride the horse - often - and the ability of the horse and rider improve over time, as well as the overall bonding.

Another thing that is being missed here are the subtle features which makeup a horse - conformation, past history, size, breed, general demeanor, etc.

There is room for great depth here which can lead to an infinite set of permutations with regard to how a horse and rider work together. The horse/rider relationship aside, all the other horse variables make the difference between a $2 horse and a $1,000 horse (Old West prices). Today's prices would be more like $500 to $100,000.

One of the points I'm stressing is that a horse's feature set can and possibly should be as rich as a complex character's, or as rich (actually more so) than the feature set of a car in a realistic racing game.

And then we move to skills. Does the horse effectively do pirouettes? Stop on a dime? Do cutting? Jump? These and more are all semi-natural activities of horses, but not entirely. I'm not asking the character to train the horse to do these things. I'm jsut saying that they don't come automatically with any and every horse.

Then we have conformation: Good features are large nostrils, a good back, balanced legs, a good stance, a proper build for its intended uses, and so on. Once again, a simple variance of these features adds depth to a player's decision making processes.

Then we have the casual things that the player does over time with or to his horse which can have positive or negative effects. I mentioned an example earlier about carrots. Other things are ill timed punishment or unwarranted punishment: expect the horse to exact revenge later.

Then we move to a horse's acute sensitivity to its environment, which can be good and bad. A horse can sense danger almost assuredly well before a human can, and if the horse's behavior is effectively modeled, and the player is attuned to this behavior, this helps the player.

[edited by - bishop_pass on December 10, 2002 10:03:02 PM]

Bishop_pass makes several excellent observations. I''ve been watching this forum and the AI forum fairly closely lately, and I think that this would be an excellent place for the two to overlap.

Horses could be implemented as artificial life. They could be bred, either by the player or simply generated with random attributes and sold at a ranch. Physical characteristics (like bishop_pass''s examples of nostril size, balanced legs, etc.) could be determined with a genome, as well as behavioral characteristics, like stubbornness, loyalty, patience, intelligence, and trainability.

This way, you could evaluate a horse by simply looking at it, seeing its teeth, its hooves, its proportions, and then by riding it, to get an idea of its virtual attitude. Different degrees of training could empower a horse to jump, or pirouette, etc. could be developed through training, with the horse''s disposition to learn each skill determined genetically.

Imbuing each animal with a unique physique, physical color pattern, and personality would deepen the experience substantially. You could buy a riding horse, to take on long trips, a racing horse, to zip across short distances for fun, profit, and survival, or a gigantic beast of burden capable of carrying enough supplies to keep the both of you alive for weeks at a time in inhospitable terrain, or to haul your heavy-armor-clad hero into battle.

Of course, the character would have to contribute a little bit to the process. You could include equestrian attributes to whatever XP system you choose to implement, and the character''s skills could be combined with the horses attributes to determine the rate of bonding and training.

Horse equipment would also become a necessity. Shoes are obvious, but saddles, bridles, harnesses, saddlebags, armor, livery, etc. could all be applied to the horse, further specializing the animal. They could be hitched in teams to wagons and carts, gradually forming a relationship with one another, and increasing their mutual productivity.

Horses might bring their own set of problems, as well. They could get sick or injured and need treatment, or run off and be killed by wolves. Wild horses could roam the plains, ready to be captured and tamed or sold.

Of course, all of these design features could be applied to dogs, falcons, snakes, or any other creature in the game universe. Wizards could have familiars, hunters could have hounds. The possibilities are nigh endless.

quote:

---

Original post by Iron Chef Carnage
Horses could be implemented as artificial life.

---

Yes and no. I wouldn't pursue that route for the sake of employing artificial life. I would only pursue that route for the sake of modeling more realistic horse behavior, if one thought it would be more effective than the actual study of horse psychology and behavior.

Admittedly, an excellent artificial life algorithm might provide an interesting 'horse tank'. But more important, is the actual study of horse psychology and behavior, which would ultimately offer the programmer the greatest insights in how to model a horse's behavior.

[edited by - bishop_pass on December 11, 2002 12:56:30 AM]

quote:

---

Original post by garconbifteck
Games have too little emotion involved in them -- I don''t think I''ve ever genuinely been scared by a game, or upset. Only happy when I get a head-shot in Halo, or something dumb and simple like that.

---

Not even an oh-shit-I-haven''t-saved-in-the-past-half-hour?

Hmm. This isn''t quite what you were going for in your attempt to model the symbiosis between horse and rider, but what if when the p.c. was in riding mode, the player controlled the _horse_? or maybe treat it like a centaur with a common center of gravity - the player is one creature that has two different types of eyes (splitscreen? top is human vision, bottom is horse vision?) a hand holding a gun, a set of horse teeth, and four hooves to play with? So you could choose to try and shoot a jackrabbit with your gun, or chase it down and stomp it to death with your hooves. You wouldn''t get real realistic horse behavior that way, but wouldn''t it be fun to play?

Well, you''re right, that isn''t exactly what I had in mind, as the horse is a separate entity with which you can bond, and suggest/command to experience all that he has to offer. Your idea has merit, but not in the context that I am thinking of.

Actually, your idea reminds me of a website I ran across where they suggested the feasibility of transplanting a human brain into a horse so the person could truly inherit the physical aspects of a horse: presumably its greater senses, speed, power, grace...

I''d consider something kind of similar to a gesture interface. Something like this:

The position of the cursor relative to the horse determines what the player wants the horse to do (direction and speed). The horse itself, depending on other variables (mood, relationship, obstacles etc) will do something which fits both its own agenda (avoiding obstacles, staying well fed, not being too exhausted) and with the player''s agenda (direction and speed). If the horse is really hungry, it might totally ignore the players instructions whereas if it is happy, fed and fit, then it will do whatever the player tells it.

This could possibly be linked with a simple punishment/reward system, so you could train the horse (don''t walk on gravel, walk on the grass instead type thing).

The other slight difference I would try is having the position of the mouse cursor just be the final destination. So if the mouse stayed in the same (absolute) position, then the horse would move over to it (eventually) and stop. This would mean that for long journeys you could zoom out and just hover the mouse at the destination. There could also be a system where the path the mouse uses influences the horse - so you could move the mouse in such a way that the horse would go in a zig zag path even if there is a straight unobstructed line between the cursor and the horse.

Trying is the first step towards failure.

Yes, but that doesn''t really address the possibilities of reigning control, which leads to the possibility of pirouettes, sidepasses, backing, jumping, etc.

In fact, the horse parameter values defined so far for the horse data structure include:
Height (at the withers)
Weight
Breed
Age
Gender: Stallion, mare, or gelding.
Intelligence
Pecking order (two numbers)
Top acheivable speed
Top quarter mile speed
Top mile and a quarter speed
Top ten mile speed
Endurance
Sure footedness
"Cow sense" factor
Coat color/type
Foot markings
Face markings
Beauty
List of friend horses
List of friend humans
List of bad experiences
List of human/bonds: (up to two)
Trotting speed
Trotting ability
Loping speed
Loping ability
Galloping speed
Galloping ability
Pirouette ability
Jumping ability
Stopping ability
Sidepass ability
Backing ability
Cutting ability
Temperment: fiery, spirited, affectionate, dull, etc.
Susceptibility to blowing objects
Susceptibility to gunfire
Susceptibility to snakes
Water entering ability
Conformation: nostril size
Conformation: base wide/narrow
Conformation: leg parallelism
Conformation: depth of chest

Many, if not most of the factors above are dynamic and change over time, and the combine with the traits of the rider to determine the final outcome, as well as factoring in the environment, events in the environment, the horse''s memory of the environment, and so on. It''s all rather deep.

I personally think you''re making this overly complicated, but...

Have you ever played _Vagrant Story_ for Playstation? It has a very interesting gamepad interface system that I think could be adapted well to controlling a horse. In VS in normal mode your character can walk any direction, jump, spin the camera angle, initiate an attack, or choose a special action from a menu. If you initiate an attack it automatically switches modes and remaps the buttons so that you can now ''chain'', i.e. do combo attacks. If you recieve an attack it automatically remaps the buttons so you can do a defensive move; when the opportunity to defend expires you are automatically returned to combat mode. If you choose a special action from the menu you do that action, then return to normal mode.

How could this be adapted for controlling a horse? Well first we could change the response to the directional pad so that up makes the horse go forward, and tapping it switches the horse to a faster gait: walk->trot->gallop. Right and left make the horse turn, and down makes the horse slow down, stop, or back up (more slowly than it walks forward). One button is still for jumping, one for the menu, one for attacking (with a gun, with a lasso, by biting, by kicking, whatever action is set as the default) and the last button ''gathers'' the horse so it can perform a more complicated move like a levade or a pirouette. Both the jump button and the gather button cause the game to shift modes and remap the buttons so that the horse can carry out the next step by carrying out an attack while in the new position, or by doing the next part of whatever combo acrobatic move the player is trying for.