(06-08-2015, 04:27 PM)Vycos Wrote: My suggestion would be:
- merge LF/HF class to HF
- give it 425 cruise speed
- the ships responsiveness to be retained so each have unique characteristic
- balance the gun slots, increase the core capacity to make them viable against VHF
- HF: pro should be: more responsive ship, 1 CD 1 CD/T slot on the bigger ships, increased number of Class 7 weapons and a core to support it across all merged HF ships, combat viable interceptor.
- HF: con should be: less hull, caped to Class 7 weapons, still slightly less fire power than VHF
I swear to god of you merge them I'll get ticked. How is that a good thing even.
(06-08-2015, 05:02 PM)Jack_Henderson Wrote: 1st: The classes have to be unified in itself. All LF should be roughly the same. They are not. Especially when it comes to size there are huge differences that make or break a LF.
2nd: Everything that is really small is near-impossible to hit. And that's no fun for either side.
Therefore:
Upscale LF to some reasonable size. The small ones are a joke.
Make all LFs somewhat hittable (like the larger LFs).
Make them less of a "1 mistake => poof" thing. When full on health, they should barely survive a nuke mine.
no I disagree leave the sizes as they are you are NERFING the ship to be worse than it already is, jack. Why is variety to you a bad thing?
Plus you are mad that you got taken out with one mine in the race
That was with NO ARMOR too... And it makes perfect sence. Before the nuke nerf even we had nukes 1 killing so it is a bad point.
Plus 500 speed on cruise is a horrible idea. Don't even get me started.
400 - 425 makes a big difference already. Why I know? I did racing before and after it was changed. One of the things you understand is that it was way harder after to have a chance to catch up. I imagine it with 500 now and wow... Everything will be broken
Don't really find any systematic in nerf/buff process towards snubs at all so far.
I suggest write a calculation program which will consider mean values and give proper output. This means that nerfing/buffing algorythm will remain one and ony while the coefficients for every class/ship are to be refined. This means that all tweaking will be set on gathering data and refining values of virtual parameters which are going to describe overall ship performance and its relation towards other classes.
(06-09-2015, 01:11 PM)Swallow Wrote: Don't really find any systematic in nerf/buff process towards snubs at all so far.
I suggest write a calculation program which will consider mean values and give proper output. This means that nerfing/buffing algorythm will remain one and ony while the coefficients for every class/ship are to be refined. This means that all tweaking will be set on gathering data and refining values of virtual parameters which are going to describe overall ship performance and its relation towards other classes.
The main hindrance to a purely mathematical approach are the ship sizes, shapes, hardpoint placement, and ease of hitting a given facing at a given angle. While the total target surface area at a given angle can be calculated (though I'm not sure precisely what tool would be used to get this data), the other factors are more challenging to express in numerical form.
(06-09-2015, 02:06 PM)Echo 7-7 Wrote: The main hindrance to a purely mathematical approach are the ship sizes, shapes, hardpoint placement, and ease of hitting a given facing at a given angle. While the total target surface area at a given angle can be calculated (though I'm not sure precisely what tool would be used to get this data), the other factors are more challenging to express in numerical form.
I suggest use bounding box of a .sur file to get x,y and z scale. Then, on one could calculate volume and center of mass of a .cmp file and use this data to generate simple similiarity formula to produce a value which could be compared or calculated further.
About turret placements - for non-tz ships one can assume the firepower output is facing straight forward, while TZing ships could have simplificated formula to calculate mean firepower at each direction, like n/6 (n - number of turrets, 6 - six sides of a cube), etc. It is being rather simple when you attempt to estimate or make a simple model for calculations. The problem comes when one needs to settle assumptions or criteria to judge the numbers. But even here we can go this way: pick up a standard situation (like, three VHFs against a GB (just a random example) which is considered to be a balanced opposition), make a system of equations, devide it into parts which consider all needed variables, and solve it to find coefficients. Put those coefficients into ships mathematical model and you will get the virtual numbers which will describe it.
(06-09-2015, 02:16 PM)Yber Wrote: You can't express ships in numbers, it's very mind limited.
Take SHFs or bombers as example. Number wise they'd win everytime.
Leave balance to the players and not to a program. Programs are stupid and have no criteria.
We are not speaking of absolute numbers but relative ones. Which means that, for example, you devide hull rating by turning speed and get N value, calculated in rather pointless HP*rad/sec, but which lets you compare two different ships.
Basics of the theory of similarity might this be, but I am not really into it.
I've tried to develop a mathematical way to approach balance once and the result was basically that an equally good, if not better, result can be achieved by using good old intuition and testing.
(06-09-2015, 02:45 PM)Tachyon Wrote: I've tried to develop a mathematical way to approach balance once and the result was basically that an equally good, if not better, result can be achieved by using good old intuition and testing.
But from what can I see now - dev team is willing to re-consider balancing of snubs, but this will lead to eventually rebalancing everything else.
To offload the ammount of work needed to be done (And how about generating .ini entries?) I suggest the first approach to be done via mathematical model, just to start with. And then - fine tuning.
Posts: 3,221
Threads: 100
Joined: May 2012
Staff roles: Balance Dev
The balance team, or at least me personally, already use plenty of objective data and "maths" for balancing. It's not just some #yolo thing where I go like "Five people complained the Eagle is overpowered, so I'm going to give it twenty less nanobots". As Tachyon noted, however, it's extremely hard to mathematically express certain aspects of the game. Shapes and sizes of ships, arguably the most important aspect of snub balance, for example, are very hard to express mathematically. I have used surface area from different perspectives (top, side, front) as well as volume in the past. While it certainly helps with balancing to have a numerical representation of a ship's hitbox, having a program decide stats based on that will not get better results than we have right now.
On the note of maths-based balancing: the class 7 and 8 gun rework that happened relatively recently was done entirely by a program, so it does happen. That's generally because guns are more straight-forward (Higher speed = better, higher damage = better, lower energy use = better. Give those aspects a certain 'weight' and you can fairly easily generate roughly balanced stats). Oh, and VHF shield bubbles have been (mostly) balanced according to volume (woo numbers) very recently.