[V2.5] How would you judge a hero's value when trading prisoners?

[stmartin]

Please share your opinion on this issue, for it's very important the AI judges hero's value in about the same way as human players.

We have planned the feature lists for HoTK V2.5. This time we will release 'Beta' versions along the way when an important feature is completed. The final release date for V2.5 is estimated to be 10.1.09.

 

[AnotherPacifist]

Obviously the more promotions the more highly valued, maybe put a gold value per promotion. There's also the matter of "fame"--little known random heroes should not be worth as much as the big players.

 

[VSPavlov]

I believe, that also:
- high-level promotions should value more, than low-level (for example, Infantry D and Cavalry D should together value less, than Infantry C);
- unique promotions should value even more.

 

[Kenjister]

I say that each hero should either have a base value assigned to them, or have a value based on the number of possible promotions. On top of those values, all current promotions should have a value assigned to each, with core promotions being worth twice as much.
Each potential Aptitude level should be worth points, with A and S being worth far more than B's and C's. D should be worth next to nothing.

And finally, the hero's loyalty should be worth ALOT. It should be the bulk of the cost of the hero. The higher the loyalty to you, the more you should have to pay.


And to clarify, are you adding a capture hero chance when you defeat a core troop? (capture as in imprison, which would render the hero useless to you)

 

[stmartin]

You will have a small chance to kill a hero in combat, then the hero has a big chance to escape, then again you have a small chance to capture the hero.

After you capture the hero, he is 'imprisoned' by default. Then you can go to the hero advisor, and choose: persuade, execute, or release at any time.

The chance for the hero to surrender will use a new formula.

 

[Kenjister]

Ooh, that sounds like an awesome system! So do they have a chance to escape as well? The only problem I can see with this is if the capture chance is too low, then we won't get to see all this cool stuff in action.

 

[Evalis]

Heroes should have some base value (dependant upon their skills, level, loyalty, and 'fame') that scales as the game continues, and the factions empire increases - the same way resource values are calculated.

The maximum value of a high level hero should be in the area of 1-2 full turns of GDP. This would probably range anywhere from 1000-5000 later game. The total value should then decrease each turn by 2% each era (so in later eras where turns are more involved the ransom ammount decreases faster) that the prisoner is held, down to a limit of 25% the initial value. The mimimum value of any hero should be in the area of 150-200 (scaling with inflation) making low level heroes much more expensive early game than they would be later game, when there are much more of them.

In addition, it would be nice to see some effect on loyalty through imprisonment. For example, each turn (with the exception of 100 loyalty heroes) the hero is imprisoned their loyalty decreases by a small ammount (1-2). Each time a hero is released via ransom, that hero should gain a moderate ammount of loyalty with the releasing faction (10-15). Maybe even something like, each time a hero is excecuted the value of that nations army increases in the calculations of the AI, while slightly reducing relations (probably some decimal point per hero).

 

[ORION11380]

I believe, that also:
- high-level promotions should value more, than low-level (for example, Infantry D and Cavalry D should together value less, than Infantry C);
- unique promotions should value even more.

I agree completely. I would suggest using an aggregated index of the hero's current and potential ability, with each value squared. D=1, C=4, B=9, A=16, and S=25

So a hero that's S/D/D/D/D would be worth 25+1+1+1+1, for 29 points.
Another hero that's B/B/B/D/D would be worth 9+9+9+1+1, also 29 points.

The first hero is extremly valuable in melee, but virtually worthless everywhere else. The second general is not as specialized as the first, but since he is so versatile, he could be valued the same.

Then each of their other promotions could be assigned an innate value, either arbitrarily assigned by the designer, or voted upon. If voted upon, I would suggest using the MEDIAN of the votes, so outliers would not skew the results.

 

[deadliver]

I would judge a hero based off of their level. Sure the promotions may make a difference between two heroes of the same level though...

 

[Evalis]

Level is not a good method of determining value. Indeed it could be argued that the =lower= the level of the hero and the more promotions he has, the more he is worth. That is to say, I would rather have a level 1 Hero than a level 10 hero if they had the same promotions.

 

[deadliver]

Level is not a good method of determining value. Indeed it could be argued that the =lower= the level of the hero and the more promotions he has, the more he is worth. That is to say, I would rather have a level 1 Hero than a level 10 hero if they had the same promotions.

I doubt this is possible. A level 1 hero having more promotions than a level 10 hero? Please provide an example.

 

[Evalis]

The level 10 vs level 1 was an arbritrary set of numbers (and a bit exaggerated). There are plenty of level 1 heroes better than level 2 or 3 heroes. Huang Zhong vs Zhang Yang is a good example, since he starts at level B in something vs C, along with 3 starting promotions instead of 1. Huang Zhong at level 2 (A Rank + 3 Promotions) is then effectively the same as Zhang Yang is at level 5 (A Rank + 3 Promotions) except that Huang Zhong only needs 3 more xp to level and Zhang Yang needs 9.

I'm actually going to take part of this back now that the whole army thing is soon to be in.. level may indeed be a relevant positive factor in determining hero worth. My bad