These forums have been archived and are now read-only.

The new forums are live and can be found at https://forums.eveonline.com/

Player Features and Ideas Discussion

 
  • Topic is locked indefinitely.
Previous page12
 

Ship Balancing - an automated approach
Author
Previous Topic Next Topic
Andy Landen
Deep Core Mining Inc.
Caldari State
#21 - 2011-10-17 22:03:21 UTC
As people learn and doctrines adapt, different ships will receive bias. Sometimes people fly a ship a lot because it is alliance doctrine, not because it is too strong. People fly ships because they are not aware that other ships are that much better, or because their skills are far from it. Popular ships are not always the best. New doctrines challenge old, popular ones, and sometimes popular ships get owned until people learn.

Better idea. CCP continues their approach of looking at the root causes of imbalance and tweaking them with careful thought and calculation for an interesting game. Can you imagine hybrids getting a 300% boost because they are 300% less popular? ouch.

"We cannot solve our problems with the same thinking we used when we created them." Albert Einstein 
betoli
The Scope
Gallente Federation
#22 - 2011-10-17 22:30:55 UTC  |  Edited by: betoli
Andy Landen wrote:
As people learn and doctrines adapt, different ships will receive bias. Sometimes people fly a ship a lot because it is alliance doctrine, not because it is too strong.


Over a range of aliances that should balance out then Big smile oh but it doesn't - the ship rankings speak for themselves, no opnion can really alter that.

Quote:

Better idea. CCP continues their approach of looking at the root causes of imbalance and tweaking them with careful thought and calculation for an interesting game. Can you imagine hybrids getting a 300% boost because they are 300% less popular? ouch.


If you'd read the proposal, you would have seen that couldn't happen. Using the stats I suggested, and your figures, the gal boat would get a net (including the slight nerf to the others) around a 0.3% boost over all (assuming it was a dps bonus), and it would get that every two months until usage fell in line. In fact even if the gal boat was 10000% less popular, the change wouldn't be any bigger, because the maximum change per cycle is fixed.

I actually think my scaling was a little conservative, it should take less than a year to close the gap ;-)

Its been 3 years since hybrids needed a balance, thats some careful calculations those CCP guys have been doing!
Andy Landen
Deep Core Mining Inc.
Caldari State
#23 - 2011-10-17 23:03:31 UTC  |  Edited by: Andy Landen
betoli wrote:


Over a range of aliances that should balance out then Big smile oh but it doesn't - the ship rankings speak for themselves, no opnion can really alter that.


The alliance part may balance out if there are enough alliances to follow every kind of doctrine with members skilled in every kind of ship to meet the doctrine, AND all alliances and people understand the merits and disadvantages of every possible doctrine. There are not enough alliances to cover doctrines, their members trained well enough to fulfill their first choice doctrines. Alliances and people do not fully understand all benefits and disadvantages to every doctrine imaginable. finally, Not every type of ship is needed the same as every other type. If fewer people fly the Lachesis than the other Recons, that does not mean that we need more Lachesis and therefore stronger eWar bonus on the Lachesis, for example.

betoli wrote:

If you'd read the proposal, you would have seen that couldn't happen. Using the stats I suggested, and your figures, the gal boat would get a net (including the slight nerf to the others) around a 0.3% boost over all (assuming it was a dps bonus), and it would get that every two months until usage fell in line. In fact even if the gal boat was 10000% less popular, the change wouldn't be any bigger, because the maximum change per cycle is fixed.

I actually think my scaling was a little conservative, it should take less than a year to close the gap ;-)

Its been 3 years since hybrids needed a balance, thats some careful calculations those CCP guys have been doing!

OK, you got me. Providing a ceiling for change can help with overcompensation.

Even still, I don't like 0.1% bonus changes, because I could never remember that. For instance, I know Rooks/Falcons and Widows get 30% ECM strength per level. This knowledge makes it much easier to know the ships and evaluate which ones I should train for, and use. Every change requires every player to learn that change. Can you imagine, "Hey fly the rook." "Sure, what is its ECM bonus?" "How should I know or remember? Think it was about 32.7% last I checked three months ago, so +/- 0.9% or so to that." Sheesh.

Additionally, training for long times with high risk is discouraged with this automated "balancing" system .. "Train for 6 months into the Widow." "OK, let me look that up. Hmm. 32.7% ECM strength, 5.3% cruise ROF, etc. OK. so in 6 months these stats can drop up to 1.8% and over the next 6 months that I am flying it, I risk another 1.8% While it could go the other way too, the risk of the ship being gimped (without an absolute floor) inhibits the investment of training long skill queues to get into advanced ships.

"We cannot solve our problems with the same thinking we used when we created them." Albert Einstein 
Fango Mango
University of Caille
Gallente Federation
#24 - 2011-10-18 00:26:16 UTC
What a brilliant proposal, I think that a lot of the people commenting here couldn't see the forest for all of the trees . . .


If I may paraphrase what the original poster was saying . . .

1) It would be good for EVE if all of the ships were a valid choice, not just a select few.

2) CCP has not done a good job of manually "balancing" all of the ships



He goes on to propose a solution that is VERY SIMPLE

1) Compare the usage rates between different ships to determine which ships are overpowered. If a ship is used more often than average it is overpowered, if it is used less often than average it is underpowerd. Every so often you nerf the "over powered" ships and boost the "under powered" ships until it reaches a steady state where they are all being equally used.

Sure there are some questions to work out (like how many people do you want flying Battleships vs Assault cruisers), and what values do you change (powergrid, shield, armor, regen???).

Once CCP came up with those rules, any 3rd year engineering student with half a brain could implement a PID loop to maintain usage balance between the ships.

And for those of you that don't see it, the neural network that does the balancing is the MASSIVE PLAYER BASE, not this program, this program just provides a control loop around the player base.

-FM
Goose99
#25 - 2011-10-18 03:44:05 UTC
This scheme worked so well in games like Starcraft 2 because they're different than Eve. It's not as simple as "ppl are building siege tanks, not archons, so nerf the former, buff the later." Eve has training time. This will create a big delay on balancing time and create rubber band effect, as people train for months towards ships based on current fotm. Despite what forum pvpers here think, Eve, and its ship usage, is overwhelmingly pve, not pvp. Basing changes on ship usage would mean balancing based on pve,which will be disastrous for pvp balance.
betoli
The Scope
Gallente Federation
#26 - 2011-10-18 10:23:28 UTC
Goose99 wrote:
This scheme worked so well in games like Starcraft 2 because they're different than Eve. It's not as simple as "ppl are building siege tanks, not archons, so nerf the former, buff the later." Eve has training time. This will create a big delay on balancing time and create rubber band effect, as people train for months towards ships based on current fotm.


As stated the rebalancing time needs to be longer than training time, thats not the rebalancing cycle time but the time it takes to level out an anomaly. After everything converged to approximately equal, changes would be tiny and not worth retraining for anyway. However if 99% of eve decides to fly the same ship even though the ships are balanced then that ship would start being nerfed - smart people won't follow the crowd, and they'll get rewarded for that (a tiny amount)

Please remember in this that nerfs and boosts are a very gradual thing, not the sudden sledgehammer associated with a one-time rebalancing exercise.

Quote:

Despite what forum pvpers here think, Eve, and its ship usage, is overwhelmingly pve, not pvp. Basing changes on ship usage would mean balancing based on pve,which will be disastrous for pvp balance.


Yup, I think thats why the OP suggested using non-high sec kills, rather than pure usage ;-)
betoli
The Scope
Gallente Federation
#27 - 2011-10-18 10:41:30 UTC  |  Edited by: betoli
Andy Landen wrote:

Additionally, training for long times with high risk is discouraged with this automated "balancing" system .. "Train for 6 months into the Widow." "OK, let me look that up. Hmm. 32.7% ECM strength, 5.3% cruise ROF, etc. OK. so in 6 months these stats can drop up to 1.8% and over the next 6 months that I am flying it, I risk another 1.8% While it could go the other way too, the risk of the ship being gimped (without an absolute floor) inhibits the investment of training long skill queues to get into advanced ships.


That would only happen (and I think the numbers used there are an exaggeration BTW) if you were training the long training because the ship was overpowered, and you had an expectation that it would remain overpowered forever. With the new scheme, you can train for a ship knowing that it will remain competitive in the longer term, sure there might be microscopic changes to the bonuses, but once everything is equilibrated, those would not be sufficient to invalidate your training decision.

In contrast, today, we have an unbalanced system in which several technologies (active armour, hybrids, yada yada) are uncompetative, and players make catastrophic decisions based on an assumption that things are roughly equal, when they aren't, or the technology falls completely out of the game (as with gal pvp boats). If CCP make a manual intervention, there will still be player rage about wasted training when it turns out in Feb that hybrids are now uber and 90% of pilots have to start over.
Previous page12