R13 stats?

[mist]

So, there's going to be a round 'later this month' (http://pirate.planetarion.com/showpo...90&postcount=5). Given the general concensus that this round's stats aren't the best ever produced i assume that there'll be some changes for next round.

So, who's producing the next set of stats? as of yet i've heard quite a few thing about who shouldn't do it, but none about who should. i feel a quick recap is in order:

the person doing them shouldn't be involved in an alliance, as doing the stats gives some form of advantage over simply looking at them - this was born out when sid did the stats and everyone complained.

the person doing them shouldn't be in pateam, as the pateam arn't capable of producing good stats. this was born out this round when they did the stats and everyone complained.

so, someone who's not in an alliance, not in pateam and who knows a load about planetarion stats is required to produce a set, asap. previous experience a definate drawback - when they last did them, odds are everyone complained.

there are also a few things to bear in mind:
the round robin approach doesn't appear to work very well, as it relies on roughly equal numbers of races, which are hard to guarentee (see r11 and everyone, shock, complaiing).

subversion is a bit of a problem. as demonstrated in r12 it's rather tricky to balance. fire subverting ships late and they're just like expensive emp. fire them early and green hostiles etc are a real problem. the combat engine won't be changing so these problems will have to be dealt with at a stats level.

also, some would like to see the return of second/third targetting (no idea if the engine can do this), while some think it would be a bad thing. pleasing both camps is preferable.

so, any suggestions as to who can produce the stats?

[yupster]

All the arguments about who should make them are all flamewars, so this is just gonna be another. I personally liked r11 a LOT more than r12, though one could argue r11 stats werent perfect either. Let experienced players help at least. (And no, not only as beta-testers cause lets face it, they dont serve much purpose :O)

[TheRat]

In my view, the community should do the stats. Heres what I would do (out of the blue, would need thinking through):

* Base the stats on those made in R11
* Run a closed beta with a selected group of people known to be competant and dedicated to do what is supposed to be done in a beta.
* Have some sort of feedback system, where each participant is required to submit an analysis of the stats after an ended "period" of beta.
* Propose an improved set of stats fromt he analysis given, and give some time for a discussion about this on a dedicated forum.
* When ready, run an open beta where the community in general can get a feel of how the stats are, and where they can suggest anything the inital crew overlooked (unlikely to be significant if the initial crew did what they were supposed to do)
* Start the round with the best set of stats PA has ever seen.

-TheRat

[Ultimate Newbie]

Well, i dont know about you, but with the number of people (especially the number of poms ) playing this game, it is highly unlikely that any ship stats will be without complaint or criticism.

Furthermore, finding someone who 'knows a load about planetarion' and who is not in an alliance is almost a contradiction in terms - everyone knows that without an alliance you are (at least to a large extent) screwed in PA.

Having said all that, i have a number of ship stats that i've been working on for some time now (Spinner and a number of others know this), and with a quick word i can modify them to suit. However, my stats do use subversion (as i invented it afaik) and i'm quite fond of it (at least in my stats). heh. However, i am a member of an alliance so i cant really be the shipstat maker person.

thumbs up to 2nd targets (though tbh with only 6 classes, 3 targets is a little too much. My stats with 10 classes however... )

[TheRat]

I would be very interested in seeing those stats UN, from a nerd point of view :-)

[jerome]

Get a "Stats Team"*, iron out the r11 set, tweak the round robin system a tad by making sure ziks don't have advantages from said subversion stuff etc. - but not by taking away that completely but by offering ways to actually defend against that etc, by just lower init for a ship etcetc.

Betas are pointless, stats, especially if we're just ironing out+improving r11 ones, won't take more than a few nights of discussion imho, all betas are to many, free speeders, eod. Just let the stats team play against each other in priv beta & have a public beta pre-signups to let everyone have a quick feel etcetc.


*get people who have already shown interest & ability - Sid, Rattie, Sovvy, Martok(whether he likes it or not:>)etc..

[noah02]

how about a beta for 4 ppl one cat one terran one zik and xan and you do about 10 different betas like this with different ppl and see what stats are to good and to bad and what ships are usless and amend them accordingly.

I know this is a stupid idea but i dont care

[Cochese]

Originally Posted by mist

the person doing them shouldn't be involved in an alliance, as doing the stats gives some form of advantage over simply looking at them - this was born out when sid did the stats and everyone complained.

You'll find maybe three people outside of alliances who are qualified to do it, and I doubt any of those three would give you the time of day, so you've shot yourself in the foot here.

Everyone complained because it was Sid doing them. Had he used a fakenick, no one would have complained. Get over it already.

I would rather see r11 stats used as a base, instead of these wank r12 stats.

so, someone who's not in an alliance, not in pateam and who knows a load about planetarion stats is required to produce a set, asap. previous experience a definate drawback - when they last did them, odds are everyone complained.

You're looking for the golden goose with these demands, and you won't find it.

Previous experience is a must, not a drawback. This "if they complained, you have failed" mentality is silly. People have, and will, ALWAYS complain to some degree or another.

When I last did a beta before the PAX beta circus, nobody really complained about the stats. Granted, we had Oreo running (read: letting us run things) the show for a few rounds, so things actually got done.

subversion is a bit of a problem. as demonstrated in r12 it's rather tricky to balance. fire subverting ships late and they're just like expensive emp. fire them early and green hostiles etc are a real problem. the combat engine won't be changing so these problems will have to be dealt with at a stats level.

This can be 'fixed' to a certain extent, but this will probably always be like playing with fire.

also, some would like to see the return of second/third targetting (no idea if the engine can do this), while some think it would be a bad thing. pleasing both camps is preferable.

Which would mean adding only secondary targets, and possibly only to certain ships--not all.

so, any suggestions as to who can produce the stats?

Plenty of people can, if you compromise your "demands" for their qualifications.

[flikman]

Why not have an open forum thread for people to their stats plan.

Then at least pateam can draw some ideas from those.

By stats plan I mean produce a list of stats, not just give generalisations.

[MAdnRisKy]

I could draw you up a list of any numnber of people who could work together to produce some playable stats. that isn't the problem and it never has been. The problem is that the people who hold the ears of those who call the shots, have historically speaking, not been crash hot in the area

the few exceptions to that rule include rnd 6 and rnd 7 and low and behold useable fresh stats.

if someone who has the ability to make things happen with the stats, wants to get in touch with me feel free
(shoots self in foot) I probably qualify for most of mist's list, except that i have had experience with doing it before. didn't hear too many complaints at the time though.

[Veedeejem!]

Finding people that are good enough to make the shipstats and that aren't in an alliance is close to impossible.
Imo it would be best if a couple of people worked on the stats together. Not all coming from the same alliance offcourse, I'm pretty sure most top alliances have atleast 1 guy capable of making shipstats, why not work together then?

[Neferti]

i think that PAteam aren't that bad at doing stats. they're bad at running effective beta tests... in my opinion, the stats at the start of the round 12 beta test were much better than the stats at the end; largely due to the fact that PAteam had a habit of listening to the people who shouted the loudest, and people who didnt really seem to know what they were talking about. For example, i seem to remember Zik being SLIGHTLY overpowered after the first stats tweak after testing began. But by the end, due to one or two people complaining ceaselessly about zik being UNDERpowered, zik was made even better.

Not saying PAteam's stats were perfect. But they could have been perfected had the beta test been run more effectively. I think TheRat has a very good suggestion in his earlier post regarding closed tests.

[Nolez]

In my opinion, I wouldn't mind seeing a rather large change in the stats. Having said that I'm going to break this post into two segments...

If you change the stats like I suggested, I wouldn't mind seeing even more races added. Yes, they would all have the same basic principal, they kill other ships, but there are plenty of ideas for specialties, while some are better than others, most of the ideas are valid. One idea, that I think is valid, is a greater effect on tech tree, roid farming and scanning based on what race you are.

Also, ships can have different specialties. Yes, we already have subbing, emp and cloaking. But there are even more ideas, such as races with weaker armor/fast eta, races the other way around, and then you can even get pretty wild with things such as races that can fake ships to throw off enemies and things like that.

Having said this, I realize that doing this would mean several things. Mainly, a lot of work, for the PATeam, and for whom ever we decide make the stats. However, I feel like having a wider variety of races would really add more strategy to the game, and allow those who really put in the time to decipher the stats to profit from their work.


As I said, this might not be possible, and as I promised, theres a second part to this post. If we are to keep the current stats, I think theres plenty that can be done.

People have suggested using the r11 stats, admittedly, I did not play, and therefore cannot speak to these stats, however I think there can still be some changes made regardless of what these stats.

Contrary to what people have previously said, I think that betas are a must, and the more the better, even if this means delaying the real round by a few weeks. Nothing bad can come out of these betas, and the more people play with the stats, the more they can learn about the stats and the more mistakes that can be found.


So basically, I wouldn't mind seeing the stats changed, and I have heard a few other people saying that this would make the game even more interesting and fun. However, undoubtedly that would have many cons and probably cannot be done. Its just my opinion that we need to really try the stats before the round so that any "kinks" can be found and we are able to really develop some good stats.

[Neferti]

nothing bad can come out of betas? what about the round 12 stats?

of course we need betas. but they need to be run properly and objectively!

It is going to be nearly impossible to find someone who:

1. Knows the PA stats well
2. Does not play PA (ie not in an alliance)
3. Is not in the PA team
4. Can actually handle the task

The stats will never be perfect and people will always complain. Maybe you should think of part of the skill of the game is being able to recognise the best race from the stats at the beginning of the round?

"Run a closed beta with a selected group of people known to be competant and dedicated to do what is supposed to be done in a beta." - The Rat

That is retarded... how is that in any way going be an accurate sample or representation of what is going to happen in the real round? I think the more realistic the beta round is the better, and someone pointed out that the PA team listens to those who shout the loudest... I totally agree, and rats suggestion is certainly not going to help that.

[Blixxard]

Originally Posted by The Master

1. Knows the PA stats well
2. Does not play PA (ie not in an alliance)
3. Is not in the PA team
4. Can actually handle the task

So. A committee.
---

Closed Beta is HC's, respective members of the community, Planetarion representatives.

Neferti your point on saying we run bad beta has some validity. There used to be a beta manager; Petru. As you can pretty much guess there isn’t one now.

I think the problem with open beta is that many people who sign up think free game and end up playing to win. The aim of open beta is to get a wider spectrum of people to iron out the bugs.

a committee is a good idea indeed... it will take longer but it will probably be more effective in creating balanced stats

Do not fully understand your point about people trying to win the beta though... isn't that a good thing since thats going to be the goal in the real round too?

[Cochese]

Originally Posted by The Master

Do not fully understand your point about people trying to win the beta though... isn't that a good thing since thats going to be the goal in the real round too?

No!

We know the game can be 'won'. That aspect is not at all in question. It would be like 'testing' the roidcap formula, when nothing has changed. The fast ticks of the Beta also make 'winning' it a different game entirely, and not indicative of how the real round will be 'won', or even how it will progress.

All it is for, is being able to launch fleets and experience combats quickly.


Remove asteroids and initiation from the beta test, and I guarantee you, people won't "play to win" anymore.

[ThunderCat]

id like to suggest webangel for the job of making them stats
dono if hed agree but he seems like the right guy for da job

[Banned]

Originally Posted by Cochese

Remove asteroids and initiation from the beta test, and I guarantee you, people won't "play to win" anymore.

For stats you might as well make the beta a battlecalc linked to the ingame combat engine.

[Envious]

Originally Posted by ThunderCat

id like to suggest webangel for the job of making them stats
dono if hed agree but he seems like the right guy for da job

based on?

[Cochese]

Originally Posted by Banned

For stats you might as well make the beta a battlecalc linked to the ingame combat engine.

Correct, but we probably won't ever see such a thing.

It makes too much sense!

[TheRat]

Originally Posted by The Master

That is retarded... how is that in any way going be an accurate sample or representation of what is going to happen in the real round? I think the more realistic the beta round is the better, and someone pointed out that the PA team listens to those who shout the loudest... I totally agree, and rats suggestion is certainly not going to help that.

a committee is a good idea indeed... it will take longer but it will probably be more effective in creating balanced stats

So first my suggestion is retarded, but when blixxard suggests almost the same thing, you agree?

My suggestion WILL help, as in the beginning you will have a serious team of beta testers who wont be trying to shout out loud. Everyone in the team will get their turn to speak.

[Tactitus]

I think the importance of beta testing for stats development is overrated. So often it seems like 'testers' are just campaigning to have their favorite/preferred race upgraded. You either have to limit the number of beta testers to people you can trust (mostly ), or else accept that a large percentage of the stats comments are going to be chaff (and hope you can spot the few grains of wheat). I think it would be more productive to have the beta testers concentrate on more general gameplay issues.

Designing stats by a committee is probably also doomed to failure. Pick a lead stats person with, at most, 1 or 2 assistants and lock them in an irc channel with a good bcalc and let them come up with some stats. If you really think you have to have more people involved, then set up multiple efforts and go with the best one.

A beta test should not yield more than a few very minor stat tweaks! Any more than that and you should probably be looking for new stats.

[MAdnRisKy]

What Tact jester the rat cochese have said!

beta testing is there to point out flaws and minor tweaks that said it is neccessary.

there does need to be a creative element with stats, which is only ever fueled by desire

using green hostiles (the very first time it was done) was probably quite a novel thing, and someone would have done it to try and exploit a weakness in someones fleet they couldn't do alone. likewise sending red defence ;-)

neccessity is often the mother of creation, and a bcalc removes these elements :/ (though i see where you're going and i agree on almost every level jester) for general creation, it's the quickest (and sadly probably the most efficient and best) method.

[Psi_K]

Just get Theam and Co to do it again.

[Alessio]

2 stats versions, that switch with eachother after 60% of the round would be cool..(60% bcz peeps have to tech up first to get BS in the first place etc) both stats availeble from the beginning of the round.. so it can be unbalanced but changes after half the round to balance it in favor of some other race.. so peeps have to make an strategical decision at the beginning of the round.. Let the PA team make nice unbalanced stats for nxt round (they are gd at that so why not let them do it)

or is that the stupidest idea ever ..?

[Conall]

On the topic of stats: This is more about the stats themsleves rather than who makes them.

The initiative spread is very small. Almost half the combat ships (17/36) fire init 3 or below. Only 25% (9/36) fire at init 5 and 6. IMO it would go along what in balancing things if PATeam used more of the init space. There is room from 1-14, with Structure killers firing at 15. If you moved structure killers to 20-24 range, or for that matter 25 like PODs, then you would have even more room for differentiation. Perhaps even have a range for a class for example FI Inits 1-4, CO 5-8, FR 9-12 and so on, with SKs and PODs having init of 25. This would make initiative similar to damage with small (read quicker) ships having better init, but doing less damage and having less armor. Perhaps you give Sub and EMP ships a class or two bump in init to give them a slight edge.

Also there needs to be balance between the three leggs of the ship stat stool: INIT; DAM/ARM and COST. SO that as one stat (INIT or ARM/DAM) is increased or decreased the cost follows suit.

I know both of these are incorporate to some degree now, but I think over the rounds keeping that balance was fallen off and attention was place more on certain areas such as damage.

[Shev]

Originally Posted by Psi_K

Just get Theam and Co to do it again.

Petru's not around any longer unfortunately - haven't seen him on IRC since November I don't think.

[furball]

Originally Posted by Conall

The initiative spread is very small. Almost half the combat ships (17/36) fire init 3 or below. Only 25% (9/36) fire at init 5 and 6. IMO it would go along what in balancing things if PATeam used more of the init space. There is room from 1-14, with Structure killers firing at 15. If you moved structure killers to 20-24 range, or for that matter 25 like PODs, then you would have even more room for differentiation. Perhaps even have a range for a class for example FI Inits 1-4, CO 5-8, FR 9-12 and so on, with SKs and PODs having init of 25. This would make initiative similar to damage with small (read quicker) ships having better init, but doing less damage and having less armor. Perhaps you give Sub and EMP ships a class or two bump in init to give them a slight edge.

The idea of only using 6 initiatives was brought in r11 I think (didn't play) after initiative had been abolished for PAX. I understand your suggestion, but I think it'd be the start of a return to the confusing days of pre-PAX when every ship had its own initiative. Personally I like how the stats work at the moment, just some tweaks are needed as opposed to totally re-working the stats.

Having 15 for structure killers and 25 for pods = artificial initiatives, nothing more, they just fire last.

As for the tweaking of the stats, I'm willing to help, although I do belong to an alliance


furball.

[Banned]

Originally Posted by Alessio

2 stats versions, that switch with eachother after 60% of the round would be cool..

That could be fun, though 60% would have to be replaced by a number of ticks, like 800 or so. Of course both sets would have to be announced well in advance, or at least approximations of the changes being made available (Harpy armor will be greatly increased, Dreadnought armor will be reduced, for example).

[Conall]

Originally Posted by Banned

That could be fun, though 60% would have to be replaced by a number of ticks, like 800 or so. Of course both sets would have to be announced well in advance, or at least approximations of the changes being made available (Harpy armor will be greatly increased, Dreadnought armor will be reduced, for example).

Imagine the chaos of attacks and defense right around tick 800 (if that was the tick). Attacks planned to land just before or just after to take advantage of the change. Of course it would be worse if they didnt announce it, launchig on a target thinking you would be ok, you cleared through defense and BAM your armor drops and their firepower increases. Think people complain now? You are right, it would be fun

[Bashar]

People complaining about r11 stats were complaining because Sid had done them, and 1up was winning, so they assumed conspiracy. It had nothing to do with the quality of the stats.

I think to be able to do the stats, all you really need is a good mathematical mind, get a rough framework and then use trial and error to sort it. The trouble (from what I have seen/been told etc.) with all the 'dodgy stats' has been that a small number of people had the major influence, and they were not necessarily people who had the maturity or 'intelligence' (note, I am not saying they were stupid, just that their brains don't work in a way compatible with doing the stats - kinda like getting a mathematical genius to do an essay on medicine).

As soon as you have one or two people who want to be the ones who 'make the stats' instead of part of the stat making team, you are doomed to failure. When you have a team doing it though, you also need a strong team leader who knows how to deal with people, and doesn't just back down to one persons demands all the time.

[Bashar]

Originally Posted by _ryzekiel_

Betas are pointless, stats, especially if we're just ironing out+improving r11 ones, won't take more than a few nights of discussion imho, all betas are to many, free speeders, eod. Just let the stats team play against each other in priv beta & have a public beta pre-signups to let everyone have a quick feel etcetc.

That is utter rubbish. The game is very complicated from a stats point of view. The stats themselves are only a bunch of numbers, which are in essence just tools. They do not dictate the game alone. Like any tools, the effects of them depend massively on how they are used. Also, anybody who has ever done anything like stats/design etc. will tell you that testing is vital. As soon as you don't test to your full capability, you are inviting trouble. If you want betas to be done properly, you get the right people in to do them, people who will take it seriously and try and test things properly, and get rid of anybody who doesn't. You also need to get people who think tactically so you can test how the stats are used.

The suggestion that testing the stats is not needed, and that a few 'prominent' people could just sit around and discuss them and sort them is.... well.... laughable. It is wishfull thinking (ummm... maybe just wishful actually, it ain't thinking). One of the most ridiculous suggestions I have seen in a long time.

[jerome]

I'm speaking from actual experience of having been part of several 'stats teams', (ie the one which has produced the stats for ******** for the past 6 orso rounds) rather than making my claim based on simple assumptions. There is hardly much you need to 'test' ingame as everything can be rather accurately deduced, I'm sure any other 'statistian' will agree with me, on this. Betas are just for confirmation of stats as most of the general refinement should be done via calculations & talks in irc-chans (etc).

The infamous stats of r6/7 were both 95% finished before they even went into betas, does that in-itself not tell you enough? Heh, all that even came out of those betas were just confirming what Petru, MAdnRisKy & co predicted beforehand. Ship Stats is all about theory, since there is no guaranteed way of finding of how the universe will turn out, in terms of racial density. Hence many educated estimations take place & considerations to various scenarios are rattled through in discussions about stats, betas will always be far too contained to be especially useful, when stats is the topic under scrutiny anyway. Only person being too wishful here is you, betas are really not that useful, unless you want what's usually ****ing obvious reiterated to you.

[Conall]

Well yes and no. Stats are very good indicators of what should happen. But there are many elements that can and will impact the outcome. So statistical analysis by definition is a prediction based on probability. Running the numbers alone is never a true picture of the outcome, only a well educated guess.

That said, the PA "universe" (and I don’t mean planets in this sense) is rather finite thus making it more predictable.

In the end the factors that reflect the overall accuracy have to do with the skill of the statistician and the variables that can affect those outcomes. In PA it is likely that a reasonable accuracy is achievable, particularly considering the dollars allocated to actually test the patterns. (Dollars being directly related to reasonable in this case)

[mist]

sitting about and thinking about them can iron out bugs, however i think testing in situ is needed to be sure.

-mist

[Tactitus]

Originally Posted by mist

sitting about and thinking about them can iron out bugs, however i think testing in situ is needed to be sure.

Stats can be "tested" with a good bcalc. In fact, you can test far more fleet combinations (assuming a reasonably robust automated test generator and test result evaluator) than a beta and possibly even a full round will ever test. There really shouldn't be any surprises.

That said, you still want to throw some live users at it; but if it falls over then it's often a problem somewhere else. In my experience in the early rounds, the weakest links were (a) stats being mistyped when entered into the game, and (b) coding errors (or undocumented behavior) in the combat engine.

[TheRat]

Originally Posted by Tactitus

Stats can be "tested" with a good bcalc. In fact, you can test far more fleet combinations (assuming a reasonably robust automated test generator and test result evaluator) than a beta and possibly even a full round will ever test. There really shouldn't be any surprises.

Agreed.

Originally Posted by Tactitus

stats being mistyped when entered into the game

Lol, hilarious

[Ultimate Newbie]

Originally Posted by TheRat

I would be very interested in seeing those stats UN, from a nerd point of view :-)

Which ones? theres about half a dozen different types to choose from . Forum mail me your email address and i can email them to you. I hope you have MS Excel .

Originally Posted by Bashar

I think to be able to do the stats, all you really need is a good mathematical mind,

Ah bugger, i'm screwed then

[Appocomaster]

I've been in the majority of the beta testing since round 5. I still have logs somewhere of when, with about 2 days left out of a week of planned betaing, the 4 race stats were brought out that was hectic. but although as many people have said most of the stats can be worked out without testing, firstly not many people can actually stare at a page of numbers and work it out, even with a battle calc, and secondly they're not going to be able to predict how other people will look at the stats and use them. And obviously it'll take time for the "big" combos to be accepted, and spread. To be fair i think more time should be given to stats, and while say r13 is running, especially near the end, round 14 stat testing starts, so that there is more time for it, even if it's not with the new round 14 code. I think having a beta team again would be fun

[jerome]

Originally Posted by Appocomaster

, firstly not many people can actually stare at a page of numbers and work it out, even with a battle calc, and secondly they're not going to be able to predict how other people will look at the stats and use them.

Then gather the people who can?

And obviously it'll take time for the "big" combos to be accepted, and spread.

Not really, no.

[Kal]

ok, 1st off I have produced a document saying how the combat engine works for the stats comitee, yes we do have one, but its rather small atm and as I have been away for 3 weeks all that has been done is my really feeble attempt to fix zik, and it was feeble flawed and the stats were really awful. So i think basing things on the round 11 stats rather than 12 is probably a good idea, which means its a good idea to get back thoose involved with them for a start.

I think its important that no single alliance has an undue influence on stats as such we did invite the alliances to help out but not many volunteered, but i'll poke them again. If anyone here thinks they would be helpful with stats pm me on irc so we can have a chat.

[noah02]

well if we use r11 stats i aint going terran (my life long choice) again may aswell go the l33t Xan 100% Xan players w00t

[Banned]

Does anyone have a copy of the round 11 stats to share with those of us that didn't play that round?

[Conall]

I am pretty sure I do on my home machine. private messageme on the boards with your email and I can send it to you.

[mist]

has anyone asked sid if he'll work on them again?

-mist

[Envious]

Originally Posted by Kal

all that has been done is my really feeble attempt to fix zik, and it was feeble flawed and the stats were really awful.

i really hope there's something done about invincible green hostiles. if you can't change that, i suggest you give Xan a deathstar class pod. yes i know there's no ship that targets deathstars. should be exactly the right thing to balance it.

[Gerbie]

a copy of the old stats
http://home.kabelfoon.nl/~hcwulff/PaXIstats.htm