Modification To The Experience System -- Thoughts Wanted

[EvilGuru]

Hi all. I am interested in what you think about my latest proposal to improve the experience system. Currently experience just affects:

• Accuracy;
• damage taken

However, I propose that it should also affect unit speed, with more experienced units having a greater speed.

Now, you might be thinking that because of the formation speed lock (a group of units only go as fast as the slowest droid in the group) that this will not be that useful. However, all that shows is that you do not think tactically enough 

When a damaged unit is retreating from the battle the formation speed lock no longer applies, and so, assuming it is experienced (are you experienced?) it will have an advantage.

But, the real advantages comes to those who use commanders. As units assigned to a commander have unified experience in that all of the units are at least as experienced as the commander. So if your commander is a very high rank your squad can have a reasonable speed advantage.

The formula is currently: speed * (1 + 0.05 * level). So if you are level 5 you get a 1.25x speed increase, or 25%.

What do you think about this proposed change? For those of you that are technically minded the code for it can be found here: https://gna.org/patch/?906 . I would also like to thank Coincident for looking it over for me (and finding a glaring error).

Regards, Freddie.

[Rman Virgil]

------->

* You may find these insights of interest Freddie

* There are 99 posts, a goodly amount quite meaty & thought-provoking (in fact, it's my "bible" in this area of development).

* This topic, evolutionary pathway & implementation domain is extremely exciting to me.

* IMHO, it holds the singular promise of making WZ even more of an extraordinary RTS - in fact, a bonafide A.R.T.S.

- Regards, RV

[EvilGuru]

An interesting read without doubt. One of the things that came up which I am quite interested in is commander AI.

I know that Pumpkin were looking into 'brains' for their units and have seen some people ask about making commanders scriptable. This would mean that there would be 4-5 pre-defined commanders scripts (written in wzscript, or depending on how it goes, lua) which would be called to control every aspect of the commander and its troops.

What do you think of this? Do you think that the current behaviour is in-effective/not flexible enough to warrant full script-ability, and if you were given the ability to script them what would you do with them?

Currently targeting (thanks to Troman) is second to none, with units being very smart about what they choose to attack. However would a script such as 'disrupt enemy oil' for a commander be worthwhile? (So it would go out look for weakly defended oil, capture it and then call in an assigned truck to defend it up).

I think it is possible to do without too much work (although even if the Queen gave her support to it I doubt it would be arriving any time soon).

Regards, Freddie.

[psychopompos]

while it is true IRL that experenced crews in lighter/less advanced tanks will mop the floor with inexperienced crews in the better equipped side.
this doesnt give the tanks greater speed...
accuracy bonus + ROF bonus would reflect RL. put damage bonus on that too and you covered the offensive stuff.

in a perfect world, units would try to evade incoming fire by moving, but that, i guess would be hard to code for, so the giving them a def bonus makes up for that.
weapons must cause "x" damage before affecting heavy bodies could be interesting too.

[EvilGuru]

weapons must cause "x" damage before affecting heavy bodies could be interesting too.

That is how the current armour system works.

If unit A fires a weapon doing 50 damage at unit B which has 60 points of kinetic armour (assuming the weapon is kinetic) then 1 damage would be done (it is 'clipped' so that always 1 damage is done). If the weapon was upgraded to do 70 damage then unit B would take 10 points of damage.

This is why it is very easy to render MG raids ineffective, if a couple of bits of research will quickly push your bases/units armour over the threshold. In effect a good player can render MGs ineffective very easily (taking them from 10 points of damage to the nominal 1).

While it is true speed is not totally realistic I would like to see one of us in a tank race an experienced tank crew. I have no doubt they would wipe the floor with us.

Shell evasion is only possible in the to-be-released 2.1 version, and only really practical for cyborgs (underrated and underused for the most part).

Regards, Freddie.

[Rman Virgil]

An interesting read without doubt. One of the things that came up which I am quite interested in is commander AI.

I know that Pumpkin were looking into 'brains' for their units and have seen some people ask about making commanders scriptable. This would mean that there would be 4-5 pre-defined commanders scripts (written in wzscript, or depending on how it goes, lua) which would be called to control every aspect of the commander and its troops.

What do you think of this? Do you think that the current behaviour is in-effective/not flexible enough to warrant full script-ability, and if you were given the ability to script them what would you do with them?

Currently targeting (thanks to Troman) is second to none, with units being very smart about what they choose to attack. However would a script such as 'disrupt enemy oil' for a commander be worthwhile? (So it would go out look for weakly defended oil, capture it and then call in an assigned truck to defend it up).

I think it is possible to do without too much work (although even if the Queen gave her support to it I doubt it would be arriving any time soon).

Regards, Freddie.

*

* You can find an explanation of this mock-up as well as what I still think (along with what Troman, Kage & Karmazilla thought at the time):

* 1st HERE

* Then HERE

- Regards, RV

[EvilGuru]

Fantastic stuff. This has been at the back of my mind all day, thinking about it now and then.

I envision such a feature panning out like so:

• The AIs would live in a .wz file;
• these .wz files would live in a command AI file.
• While in-game the user would be able to select a commander and then use a combo-box like widget to select their desired AI;
• The AI would take immediate affect and should be seamless to the user (that means no sub-menus or dialogs with extra configuration options).
• The script would do its thing until either the commander dies or the AI is removed/changed.

I am currently tempted to say that the AIs will not be changeable by the user, if one wants a different set then they will need to mod the game. This is so to not give programmers an unfair advantage. Everyone plays with the same available set. By not making them editable it also means that research topics can be added for them. Although this is only a temptation, others input would be useful.

What one needs to decide on next is what hooks to give the scripts. So when do we ask the script for an opinion on something. There are the obvious ones such as when we order the commander to do something, add/remove a unit or it destroys a unit. But if you were a commander on the battle field (imagine your sole job in the world of WZ was to control a group of units) when would you like to be consulted?

Currently this is all just idle thought, however substance will soon follow if others share the same thoughts.

Regards, Freddie.

[Rman Virgil]

------->

* Those are keen insights EG.

* My most up-to-date thinking on this goes to what I am seeing as an essential intermediary stage.

* One of my team mates brought it to my attention like so:

The "Mini/Mega Map":

Some ideas I had while you were all talking about it led me to think of the Command Relay map in Tribes 2.

I'm sure a few of you are scratching your heads as to what I mean, but in that game they had this really neat Map Overview function where players could create nav points, give bots/players orders.

The map of course was from a satellite view, but the sensor capture of play/bots and enemy equipment/vehicles/players ect. was very well made. Not that the ai ever listened to you.

My thought was that the new "Mega Map" could function very much like that.

* To which I responded:

Your comments on the "Mega-Map" & connecting its functionality to the "Tribes 2" Command Relay has given me an additional insight into the GCI - that is, as a terrific intermediary stage to advancing Command & Control such as to make fielding 2-3 command groups with different vectors & velocities in the same theater of operations 100% more doable than it is now....

This, for me at least, represents a major breakthrough in advancing tactical MP to the richness of RL military maneuver & combined arms engagements. Really sweet.

- Cheers, RV

[Rman Virgil]

------->

* I should make a key clarification.

* My team mate mentioned a "SAT-View" tied to the "Mega Map" & the "GCI" had a "World Map" toggle feature whose implementation is not explained.

* Well the implementation is NOT a "Sat-View".

* HOW the "Mega-Map" data is achieved (or populated) is thru Commander deployed "M.A.V.s" (Micro Airborn-Recon Vehicles or Drones)....

* The only drawback of the moment is the animations that the engine can currently support. Frankly, in this instance, I think there's viable options so the implementation shouldn't have to be stymied on the horns of a necessary overhaul (like what is posed by any effort to change the Oil Derrick resource or making a "Jump" chassis / propulsion like the BatMobile in the last movie, "Batman: Begins" - originally a military vehicle to ford rivers or jump fortification walls, for example).

- Regards, RV

[Coincident]

We're going quite off-topic here, but as for the wz minimap there's some serious issues with the map on very large maps (where you need to scroll the minimap to see all the map). I don't know if anyone has changed that yet in the trunk version, but on 2.0.8 you're trying to move your screen to some point near the edge of the minimap, and instead end up scrolling the minimap (and your screen) at ultra-fast speed to some black corner of the map.
I think it would be cool to be able to zoom out the minimap, enough to cover the whole map all the time.



Now back on topic, I've told this to EvilGuru in IRC before and I'll say it here again:
I think it's a GREAT idea to have units move faster as their rank goes up, because they can all move faster as a whole when under a commander, and most of all because they will retreat and go for repair faster.
Now you might be thinking: "What the hell do I care if they retreat faster if I only want to crush my enemy?"
But if you're having a hard game and your enemy is kicking your ass, moving to repair faster means the unit survives longer, so it can kill even more units later, get higher rank, move even faster, survive longer, and so on...
And also don't worry about high-ranked VTOLs, as they will never move faster then 700 speed (max speed of the propulsion).

So about this one, thumbs up for EvilGuru. 



Finally, about the commanders:

However would a script such as 'disrupt enemy oil' for a commander be worthwhile? (So it would go out look for weakly defended oil, capture it and then call in an assigned truck to defend it up).

Well it's obvious that commanders need to be more useful in the future, but I don't agree you should make a commander that would actually play the game on his own; not to the point where it decides which enemy locations to attack and go there without the player's consent/order (even if the commander is 100% sure they are undefended).

I think the player should order the commander to go to one spot, and it should be able to defend that spot (or pursue the enemy) in the best possible way. But even if the commander and it's units are idle for 30minutes (while the player is busy), it should never decide what to do next and move away, without an order from the player.
The commander should be responsible for the low-level tactics. High-level strategy is for the player to govern.


One thing that could be implemented on a commander would be something like: target weakest units first.
That's what I usually do when I'm controlling the units myself:
If there's a weak enemy near my units I try to kill it first, and leave the heavy, bulky, armored tanks for last.
If there's a damaged enemy near my units kill it first, and leave the enemy at full health for last.

This means the weak enemy units will die faster and not shoot at me anymore when I take the time to kill their strong units later.


To decide which enemies are weaker (armor-wise) the commander could also determine which weapons are under his control and which enemies are more vulnerable to them.
For example:
- The commander has mostly machinegun units. It should target infantry first.
- The commander has mostly lancers/tank-killers/scourge-missiles. It should target tanks first.
- The commander has mostly flamer units. It should target bunkers first.
etc.



That's just my opinion, sorry for the long post.  :-X

[Rman Virgil]

We're going quite off-topic here.....

* Wow.... you really believe that ?

* Hmmmm.... well then, if you wanna know why that assessment is in error.... let me know and I'll break it down for you.

* Also - In NO way shape or form are we suggesting, or ever have, "Autonomous Commanders". In fact all proposals years ago recognised that as a crucial caveat - that is, something to be AVOIDED. And if you actually read the referenced "Exp-Rank" &  "GCI" threads you'll see the Player-General makes ALL Macro-Decisions.... and that what we are dealing with is having to "Baby-Sit" units / combat groups so they don't commit STUPID, Obvious, acts of dumb-ass suicide as well as creating rewards for unit presevation-persistence to discourage the "canon-fodder" mind-set of mindless swarm "tacs" - among a handful of other game design goals.

- Regards, RV

[EvilGuru]

Yes, the commander scripts will hopefully make commanders seem less pre-programmed. So they will be adaptive. The scripts will be able to do like Coincident suggested and look at what units make up the group and assign the most effective targets based off of this.

What additional functionality will be available I am as yet unsure. I am quite confident that they will be able to dynamically re-write the build queue for any factories which they are assigned will be available. So if you commander has two bunker buster units assigned and some pesky VTOLs take them out and it just so happens that your factory assigned to the commander has some bunker busters queued after cannon tanks it would be able to move the bunker busters to the front of the queue.

This gives commanders enough control to make an influence but not so much as to allow them to empty your bank account.

I think it would be cool to be able to zoom out the minimap, enough to cover the whole map all the time.

We have this in trunk, I am not sure about 2.0.x. I believe you need to mouse-over the mini map and then use the mouse wheel to zoom it in or out. Per has also added some more modes for the mini-map in order to see what ones people like the most. It should be a lot more effective and useful in 2.1 than in 2.0.x.

Regards, Freddie.

[Rman Virgil]

* That's all promising & kwel.... definitely something to look foward to.

- RV

[Rman Virgil]

* Wow.... you really believe that ?

* Hmmmm.... well then, if you wanna know why that assessment is in error.... let me know and I'll break it down for you.

* That came off a bit like an a-hole.... excuse for that tone.

* I do consider "Exp & Rank" enhancements joined at the hip to "Command & Control" mechanics and as such the "Mini-Mega Map" implementation as described above is crux (as part of "Comand & Control") & therefore very-much On-Topic.

* If you create a System Dynamics Modeling Analysis of the interacting feed-back loops of the GPMs here involved (which I do as a designer) it becames plain as day.

* But if you wish to keep them "discrete" then that is your perogative. If you run a "discrete" implementation it'll become apparent & I guarantee you'll end-up going back to re-address the essential connection between the two. I was just giving you a friendly heads -up, take it or leave it, it has no impact on what I'm doing WZ-wise.

- regards, rv

[DFStormbringer]

""While it is true speed is not totally realistic I would like to see one of us in a tank race an experienced tank crew. I have no doubt they would wipe the floor with us.""

^^^^
this while placed  from the right spot of the heart made me laugh
yes.. said THAT way..  speed boost seems logical but your comparison is a tad off..
logistically being its a combat game..  and the comparison would be a greenhorn combat unit vs a veteran one...
take that SAME race..  2 identical tanks.. one manned by fully trained but non war experienced combat personell.. and pit them against a veteran crew thats been overseans in combat for over a year..  and race them.. unless theres some serious terrain challenges to the course.. the results of the race would still be up in the air for either side.

the ONLY difference wold be something yo MIGHT be able to actually impliment in game... elevation resistance.
though not really used as a tactical tool.. terrain elevation and  moving acrost slopes and gradients does radically change the units movement speed..  a greenhorn would take their time and go alot slower over said obstacle then a veteran thats learned the ins and outs of their vehicle and thus kow its limitaions casing them to go through rough terrain alot faster.

accuracy boost..  i said it before in another post...  sure. on paper it sounds great.. but unfortuantly by the time you chew through most the weapons upgrades.. even the greenhorns have accuracies well into the 90%.. and wiht hard hitting weapons only needing 3-4 hits to kill a unit.. then that 10% really dont make much of a difference.

added damage and armor to balance for evasion.. both of these are solid ideas.. even extra hit points...
and being max level is so hard to come by in most games.. how about making the hero class units even gain special abilities like double rate of fire..  added range because of weapon familiarity.. or even double auto repair speed.. or thier weapons getting an added explosive area of efect blast radius??

may seem like alot to tweak on the hero unit.. but realize that it takes over 500 kills to reach that hero status.. thats ALOT of grinding for one lil unit.. and my BIGGEST game ever barely hit 3000 kills vs 7 AI players and thats only because i wanted to farm the entire tech tree first...
even that that game would have produced MAX.. 6 hero units..  most standard games being a MIRACLE to pull off even ONE hero unit.. as ive NEVER managed to achieve one before..  so giving some lil impressive boost to the hero would really make for not only gloating rights but a worthy strategy point but not enough of a boost that its going to go out and by itself destroy the enemy single handedly.. but enough to know once the enemy sees it grouped wiht a squad that the enemys in for a challenge.