Experience stats from 250+ replays - Page 4

Experience stats from 250+ replays

Classic game design, maintained to please you...

Moderator: Content Developer

User avatar
tyran_nick
Posts: 47
Joined: 07 Mar 2010, 12:42

Re: Experience stats from 250+ replays

Post by tyran_nick »

This is quite interesting. I don't see why we cannot model these data to provide a robust balancing mechanism for building new units. I presume that given particular type of unit, weapons, etc we could have an equation to provide a suggestion for resources that modders could choose to follow.
Potentially it should give a total sum of resources and a mod specific conversion factor between resources (I pressume they will be collinear from these data, which will make the actual estimation tricky).

Controlling for player experience, unit type (vehicle, plane, etc) and game, balanced units should converge on average to the same experience, shouldn't they?

I might have missed it, but it seems as if how much beating a unit can take seems to be irrelevant?
User avatar
knorke
Posts: 7971
Joined: 22 Feb 2006, 01:02

Re: Experience stats from 250+ replays

Post by knorke »

I don't see why we cannot model these data to provide a robust balancing mechanism for building new units. I presume that given particular type of unit, weapons, etc we could have an equation to provide a suggestion for resources that modders could choose to follow.
It is interessting indeeded but you can not base game balance on that for a number of reasons.
Not for BA and even less for other games.
balanced units should converge on average to the same experience, shouldn't they?
No, it does ignore too many other factors. ie T2 might be more effective but still be balanced because of the costs of the lab. If anti air defense makes the enemy not even try to attack by air it has also fullfilled its role, but that effect is not measurable in unit xp.
Anyway, the average of experience is not even that important. It is more meaningfull how and when the experience is gained* because it tells how long a unit stays usefull (ie if T1 gets useless over time or not) and how "risky" a unit is to make.
If every unit of a unittype is relatively safe to gain medium amount of XP that is different to a unittype that either more of a gamble and either totally pwns but may also be a total waste.
ie a tank is mostly a pretty safe investment. But a nuke can either be a game winner or get destroyed without doing anything damage at all, if the enemy had an antinuke ready.
So basically "balance formulas" are fail.

*would be easier to tell with color gradients (http://springrts.com/phpbb/viewtopic.ph ... 92&start=5)
Saktoth
Zero-K Developer
Posts: 2665
Joined: 28 Nov 2006, 13:22

Re: Experience stats from 250+ replays

Post by Saktoth »

Does this factor in self-d? That will make any player who ragequits or self-d's at the end of the game make all his units count, while a player who is winning will have a bunch of units that dont count. That is going to skew your stats, though I dont know in which direction.

Anyway, we have been doing this a long time in Zero-K and among other confounding factors is the fact that you are very likely to just sacrifice huge numbers of units to break an enemy defense if you are winning. Pouring raiders into a base just because they are cheap, fast, and will get there more quickly and end the game. This sort of empirical data is a valuable tool in confirming or refuting your suspicions though.
Manmax
Posts: 78
Joined: 19 May 2011, 13:57

Re: Experience stats from 250+ replays

Post by Manmax »

What effect does the growth of a unit's xp have wrt its behaviours or skills?
User avatar
knorke
Posts: 7971
Joined: 22 Feb 2006, 01:02

Re: Experience stats from 250+ replays

Post by knorke »

The engine supports increasing reloadSpeed, maxHealth, health and power when a unit gains experience.
(afaik the increased power makes other units more likely target to target the experienced unit)

experience is also used in various places of weapon/targeting code ie it makes units aim more accurately at moving targets. Some weapons get more accurate (ie by decrease of sprayangle)

I assume BA uses all those but not sure.
I think usually only the health increase matters somewhat.
Pako
Posts: 174
Joined: 12 Jul 2009, 18:57

Re: Experience stats from 250+ replays

Post by Pako »

I think currently in BA experience is way too ineffective mostly unnoticeable but would it make the game better if it would be a major part of the gameplay?
User avatar
knorke
Posts: 7971
Joined: 22 Feb 2006, 01:02

Re: Experience stats from 250+ replays

Post by knorke »

This isnt really about the effect of experience on gameplay.
I guess experience was just used because from its value its easy to see how effective a unit was, without having to log the damage done by each unit.
Manmax
Posts: 78
Joined: 19 May 2011, 13:57

Re: Experience stats from 250+ replays

Post by Manmax »

increasing the role of xp could be an alternative to auto heal for encouraging people to spare their units and better micro-manageg them.
Post Reply

Return to “Balanced Annihilation”