PDA

View Full Version : Combat Balancing Dev Diary


Archived Post
11-09-2009, 10:26 AM
Ten Ton Hammer's posted the second part of Systems Designer Brian "Balseraph" Urbanek's developer diary (http://www.tentonhammer.com/node/76412) on their site today. In it, he discusses the mechanics of combat and how to balance the various systems in the game around it. It's definitely an enlightening read, full of difficult maths and crazy stories of love and loss. Enjoy!


Link to the news article. (http://www.champions-online.com/node/594666)

Archived Post
11-09-2009, 11:52 AM
Can you tell Balseraph I love him and want his babies (braised in a nice orange creme sauce) for writing this up? Since I didn't get to be a game developer, this is like living vicariously...

Archived Post
11-09-2009, 11:54 AM
Wow, big thanks to Brian Urbanek for writing these dev diaries! I find it really helpful to know about the target baselines for damage, hit points, etc so I can tell which of my chars are above average, and which are below average.

Quite often the math used in videogames is extremely simplified, even in big production RPGs, so that all kinds of levels of exploiting are possible that were never considered by the devs. I like to see all the thinking that is behind the arithmetic models for CO because it helps me anticipate and understand the changes that are going on.

For example, I have a lvl 22 character that was effected by the Ego blades / incisive wit nerf, and because I can't make that character look the way I want them to (two ego blades and a shadow form) anymore, I don't play them anymore. For me, this is the most painful change to CO since head start, but at least now that I see the arithmetic involved, I see why the devs decided it had to be done, which lessens the sting of the nerf.

At the end of the article Brian writes:

Well, I have about seventeen more chapters that I could write about all of this, but I'm already about three times my target word count, so I'll just cut it off here with a big TO BE CONTINUED!

Thank you so very much for writing as much as you did, and I look forward to reading all the future installments of Brian's dev diaries. A minor suggestion I have to offer is to possibly make them video-diaries if that would save time for Brian instead of written articles.

Archived Post
11-09-2009, 11:55 AM
Thank you so very much for writing as much as you did, and I look forward to reading all the future installments of Brian's dev diaries. A minor suggestion I have to offer is to possibly make them video-diaries if that would save time for Brian instead of written articles.

Ew, no, writing > speaking for this. Too many numbers/equations to comfortably do with speech. Besides, have you ever seen computer people talk?

Archived Post
11-09-2009, 01:25 PM
It's kind of disturbing to see statements that are manifestly not true in these articles. For example: The result of that is pretty simple: no power in the game can (or, at least, none SHOULD...) have an energy cost over 100, and energy costs don't increase as you level up..

There are no powers in the game which cannot be activated for 100 energy, but there are many powers which cannot be used in an effective manner without well over 100 energy (or large quantities of energy cost reduction). For example, sustained powers with escalating damage are usually a bad deal unless you can run them for a full maintain, which will often result in a base energy requirement in the 200 range. Running power armor toggles is pretty much futile without a lot of energy.

Archived Post
11-09-2009, 03:43 PM
It's kind of disturbing to see statements that are manifestly not true in these articles. For example: The result of that is pretty simple: no power in the game can (or, at least, none SHOULD...) have an energy cost over 100, and energy costs don't increase as you level up..

There are no powers in the game which cannot be activated for 100 energy, but there are many powers which cannot be used in an effective manner without well over 100 energy (or large quantities of energy cost reduction). For example, sustained powers with escalating damage are usually a bad deal unless you can run them for a full maintain, which will often result in a base energy requirement in the 200 range. Running power armor toggles is pretty much futile without a lot of energy.

This is why I think items that give 50% energy discount to melee, aoe, range is "broken" because huge energy cost for Force Casacade is there to balance that "heavy" charged up damage.

I think those 50% items should also come with damage penalty to restore that Damage/Energy ratio equation.

Yeah, I was a bit puzzled when I first read "no power in the game can have an energy cost over 100". I still enjoy the article though.

Archived Post
11-10-2009, 07:57 AM
It's kind of disturbing to see statements that are manifestly not true in these articles. For example: The result of that is pretty simple: no power in the game can (or, at least, none SHOULD...) have an energy cost over 100, and energy costs don't increase as you level up..

There are no powers in the game which cannot be activated for 100 energy, but there are many powers which cannot be used in an effective manner without well over 100 energy (or large quantities of energy cost reduction). For example, sustained powers with escalating damage are usually a bad deal unless you can run them for a full maintain, which will often result in a base energy requirement in the 200 range. Running power armor toggles is pretty much futile without a lot of energy.

I kinda have to agree with this. i love the artcles but I did find the statement Pantagruel quoted a bit off. Sorcery alone has powers ou HAVE to charge to cast. These are well above 100 energy for their base cost. You can say, "well if you went with the recommended stats and went int you would have no problem with this" but the key word is that they are recommended stats. If no power should have over 100 energy then you shouldn't need over 100 energy to throw down a circle or summon a pet, or drop down sigils.

Archived Post
11-10-2009, 08:20 PM
I'm loving these articles. It occurred to me that like most great math systems if you put junk in you'll get junk out. Therefore, it's imperative that testing is performed to determine if performance is meeting the metrics established in the design. In addition, it's imperative that customers provide feedback to development of performance.

How was the 20 second life span really estimated? This is clearly a metric, hopefully specified in the scope of the project, that is related to customer satisfaction and feel of the game. In some cases, metrics are determined by reverse engineering similar systems. In other cases they are determined using logic and gut feelings. Was this metric determine in either of those two methods or in a different method?

I am still curious to find out if a quality chart (X-bar and R) could be used to log DpS and identify if the system would be considered in control and if there are outliers or points beyond the control limits what significant factors are causing those points to be over or under the limit. I'm also wondering if an X-bar and R chart would be the appropriate quality tool to use because there should be n! different combinations of powers (i.e. different machines producing a product) and if they were all combined it would lead to stratification and too much noise would enter the system to make a reliable chart.

Thanks again for the article.

Archived Post
11-10-2009, 09:33 PM
I am still curious to find out if a quality chart (X-bar and R) could be used to log DpS and identify if the system would be considered in control and if there are outliers or points beyond the control limits what significant factors are causing those points to be over or under the limit.
You've lost me here, but it sounds rather interesting.

What are X-bar and R charts, and what are there advantages over other methods?

Archived Post
11-11-2009, 07:09 AM
...factor in X, divide by Y, subtract the price of rice in China... and it all works out to about 5,787.656 hp as the target value at level 40 by which our average player will survive for exactly 20.0 seconds.

So this is why my HP seems to be off, it's fluctuating to the market price of Chinese Rice, eh. Time to invest in HP! To the commodity trade!

Archived Post
11-13-2009, 07:31 AM
You've lost me here, but it sounds rather interesting.

What are X-bar and R charts, and what are there advantages over other methods?

X-bar and R charts are a tool used in SPC (statistical process control). A Google search turned up a decent description at the following web site: http://www.qualityamerica.com/knowledgecente/knowctrXBAR_RANGE_CHARTS.htm. Like I said in the post I dont' think it would be the best tool to use because each power set combination would be considered a different machine. If the same measurement is taken from different machines then there will most likely be too much noise in the system and the result will be a useless and misleading chart.

What other tools are available? Well right now I'm not aware of any, however, that does not mean some do not exist. I would tend to lean towards the basic statistics tools of mean and standard deviation when measuring a single variable that is produced by several different power builds. However, if you run an X-bar and R chart for a single character then you should be able to see statistically how the character grows with each level and power change. In addition you should be able to determine if the change is significant and how significant it is.

Archived Post
11-17-2009, 06:58 PM
Cool. Thx for the follow-up post.