• Play
  • About
  • News
  • Forums
  • Yppedia
  • Help
Welcome Guest   | Login
  Index  | Recent Threads  | Register  | Search  | Help  | RSS feeds  | View Unanswered Threads  
  Search  


Quick Go »
Thread Status: Normal
Total posts in this thread: 106
Posts: 106   Pages: 4   [ First Page | 1 2 3 4 | Next Page | Last Page]
[Add To My Favorites] [Watch this Thread] [Post new Thread]
Author
Previous Thread This topic has been viewed 23933 times and has 105 replies Next Thread
Shuranthae

Member's Avatar


Joined: Apr 1, 2004
Posts: 9758
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

...if the percentages are based on population and not the numbers behind the scores, then why is it that different puzzles have different numbers of Ultimates?
[Jul 27, 2007 1:50:17 PM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
sashamorning

Member's Avatar


Joined: Jan 19, 2006
Posts: 3634
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

 
...if the percentages are based on population and not the numbers behind the scores, then why is it that different puzzles have different numbers of Ultimates?

Because it's based on the number of different pirates that have played that particular puzzle within the last 10 days.

~Sasha
----------------------------------------
Sashamorning - Fiery-haired, fiery-tempered, 100% me.
~~~~~~~~~~~~
Shakespeare wrote: 
This above all: to thine own self be true,
And it must follow, as the night the day,
Thou canst not then be false to any man.

[Jul 27, 2007 1:51:51 PM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
Shuranthae

Member's Avatar


Joined: Apr 1, 2004
Posts: 9758
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

Cleaver wrote: 
Shuranthae wrote: 
Question: Is activity based on all of an active character's standing or just the standings of a character that aren't dormant? I.e., if all I do is gun then only my gunning standing factors into everyone else's standing and all my other standings are ignored.

No. If you play, all your standings are counted. If you stop playing, your standings aren't counted.

[Jul 27, 2007 1:56:25 PM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
DreadedChris



Joined: Jan 27, 2006
Posts: 3334
Status: Offline

Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

Shuranthae, Liz just -directly- contradicted that and said explicitly that it was on a per-puzzle basis. In this thread.
----------------------------------------
~< The Dread Pirate Chris >~
[Jul 27, 2007 2:01:33 PM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
sashamorning

Member's Avatar


Joined: Jan 19, 2006
Posts: 3634
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

Cleaver wrote: 
Shuranthae wrote: 
Question: Is activity based on all of an active character's standing or just the standings of a character that aren't dormant? I.e., if all I do is gun then only my gunning standing factors into everyone else's standing and all my other standings are ignored.

No. If you play, all your standings are counted. If you stop playing, your standings aren't counted.

That doesn't seem to fit with what Liz just said, then. I would tend to think that whatever Liz is saying is the way it currently is, but feel free to correct me.
 
if either a.) there are less than 100 people who have played a puzzle in past 10 days (*cough*ice*cough*), then you won't see ults. If people are too bunched at the top such that it will be impossible to set a cutoff such that only 1% of people exceed it, it falls back and there will be only legendaries.

It looks like bnav is just based on the standings of those players who've played in the past 10 days. Maybe the standings formula has been tweaked since you asked the question. When was that?

~Sasha
----------------------------------------
Sashamorning - Fiery-haired, fiery-tempered, 100% me.
~~~~~~~~~~~~
Shakespeare wrote: 
This above all: to thine own self be true,
And it must follow, as the night the day,
Thou canst not then be false to any man.

----------------------------------------
[Edit 1 times, last edit by sashamorning at Jul 27, 2007 2:04:50 PM]
[Jul 27, 2007 2:03:48 PM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
t0rqt

Member's Avatar


Joined: Aug 28, 2005
Posts: 255
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

hrm, so if i don't do crafts for a week, but i'm active, my skillz expire.
but, if i'm active, my other shipboard activities don't ?

so, i may have hit ult in sailing, last month,
not sailed at all, cuz i'm going for a crab trophy,
and still be on the ult list for sailing?

boo.
----------------------------------------
erro : karma (mid&mala) - bad company (hun)
 
artemis: ...most people voted for "ur mom"

[Jul 27, 2007 2:10:09 PM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
Shuranthae

Member's Avatar


Joined: Apr 1, 2004
Posts: 9758
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

Well then I am most curious as to which of these Ringers are wrong. Liz, can you check?
[Jul 27, 2007 2:11:25 PM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
sashamorning

Member's Avatar


Joined: Jan 19, 2006
Posts: 3634
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

I wrote: 
I would tend to think that whatever Liz is saying is the way it currently is
...
Maybe the standings formula has been tweaked since you asked the question. When was that?

----------------------------------------
Sashamorning - Fiery-haired, fiery-tempered, 100% me.
~~~~~~~~~~~~
Shakespeare wrote: 
This above all: to thine own self be true,
And it must follow, as the night the day,
Thou canst not then be false to any man.

[Jul 27, 2007 2:18:22 PM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
Shuranthae

Member's Avatar


Joined: Apr 1, 2004
Posts: 9758
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

14 months ago, when they were reworking the rating system a bit.

It still seems kind of off if it's based on population... didn't Chavez do some math regarding that?
[Jul 27, 2007 2:42:44 PM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
Corwyn

Member's Avatar


Joined: Mar 12, 2005
Posts: 117
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

Personally I think another problem that goes along with the argument that the stats are skewed because people who are bad at a puzzle wont play it, is the knowledge among those people that getting low scores now will make it hugely more difficult to get a high ranking later. This encourages people to avoid doing a puzzle unless they know what they are doing, rather than learning it. Now I don't know how long the "history" of your stats is kept, but it seems too long? Punishment for trying to learn a puzzle seems like a bad thing to me.
----------------------------------------
Corwyn on the Meridian Ocean


[Jul 27, 2007 5:15:12 PM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
TheRack

Member's Avatar


Joined: Jan 5, 2005
Posts: 6116
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

 
And I apologize to TheRack if my earlier comment was taken to be demeaning. I, possibly falsely, interpreted this as a nonserious post meant to be taken as a joke. In my defense, it was 4 in the morning.

None taken. Truth be told, whilst it was intended to be a serious post, my initial post was very vague and poorly put together...

Edit: Damn it... I'm sure I hit the PM button... Don't mind me...
----------------------------------------
Cephalopod, on poker, wrote: 
I've said it before, and I'll say it again: it isn't rigged.

Period. End of story.

----------------------------------------
[Edit 1 times, last edit by TheRack at Jul 27, 2007 8:31:42 PM]
[Jul 27, 2007 8:27:05 PM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
Rick9109

Member's Avatar


Joined: Sep 19, 2003
Posts: 6716
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

Performing in the top 10% of puzzles may not actually mean you go up quickly, since some of the puzzles try to force you to play a certain way at certain star levels.

A good example is bilging where the exact same move that will be a giant score at ultimate or legendary is actually harmful to your score at master/low-renowned, due to having so few moves to be "efficent." They try to force you to cascade-but the problem with that you hit a certain point cascading quickly become s an ineffective strategy.

This is way different from sailing where a quick donkey or a bingo is always a good move. As much as I have a problem with the legendary puzzles in sailing, at least you don't have to clear singles to get up to that level.
----------------------------------------
Rome
Pirates of the Damned, Crimson Tide.
I don't care what it did to them, the game's been good to me.
[Jul 27, 2007 9:52:54 PM] Show Printable Version of Post        Send Private Message    rick9109    cactusrome [Link]  Go to top 
Chavez67

Member's Avatar


Joined: Aug 26, 2003
Posts: 7739
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

dreadedchris wrote: 
Suppose that each puzzle adjust your standing such that,
standing' = standing + ( ( score - standing ) / 10 )


That's not how it works. We have definitive confirmation that it's a average (possibly weighted) of the last X scores, not a running adjustment to a single number. I can't find it right now, but search around a bit, try the wiki first.

liz wrote: 
Midnight certainly has >> 100 active bnavvers. Debugging...


Yeah, I told you guys there's no way it worked like that. I could maybe see how more than 1% of the active distillers on an ocean could all score CC16, but there's no way that 1% of the active battlenavvers all have exactly the same score.

dominator wrote: 
See, Liz, this is why we love you. Not only do you tell us how bits and pieces of things work, you go and fix them! And right away, too!


Speak for yourself. I love Liz because she's a great admiral, solid blockade navver, used to pvp dockpressers back in the day, and she's pretty hawt to boot. That she does code is incidental. (plus... there's no props for 'promptness' in addressing the lack of ultimates ... :P )

shur wrote: 
It still seems kind of off if it's based on population... didn't Chavez do some math regarding that?


I linked some pretty pictures and spoke some mumbo jumbo, clearly not math, just allusions to it. In any case, what we've definitively discovered so far is only that our information in how the standings system works is lacking, and Liz appears to be on a mission to enlighten us, which is cool by me. Makes this thread a gem, in my opinion, and it's not even done yet.
----------------------------------------
 
 
 
I did it.
clarify
It was me.

[Jul 27, 2007 10:23:36 PM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
abacadafa

Member's Avatar


Joined: Jun 13, 2006
Posts: 493
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

 

I could maybe see how more than 1% of the active distillers on an ocean could all score CC16, but there's no way that 1% of the active battlenavvers all have exactly the same score.
(italics added)

Well, that's the problem with distilling, right there. omghax!1!!!11!
----------------------------------------
.
[Jul 28, 2007 7:10:26 AM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
Benzene265

Member's Avatar


Joined: Jul 18, 2005
Posts: 6036
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

Hey, I've been booching my labor alts off the SW list. It's actually harder for me to play badly than it is for me to play well. It's kind of an interesting challenge, even if it lacks the Vegas music. Do dado da dodado Dooooooooo!

Edit: The real solution for the crafting puzzle curve is to kill labor alting. Try to make it so that I can only use Rhodin. Adding all the crafting puzzles is certainly a step in the right direction. There was a thread about getting rid of offline labor, but I have no idea where it is.
----------------------------------------
A Ghyslaine and a Rhodin for every Ocean, but mostly on Viridian.
Make the natural choice for our oceans: Google Rhodin Blonde!
----------------------------------------
[Edit 1 times, last edit by Benzene265 at Jul 28, 2007 6:02:16 PM]
[Jul 28, 2007 5:56:56 PM] Show Printable Version of Post        Send Private Message    Have you considered becoming a well-tanned blonde? [Link]  Go to top 
DreadedChris



Joined: Jan 27, 2006
Posts: 3334
Status: Offline

Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

 
dreadedchris wrote: 
Suppose that each puzzle adjust your standing such that,
standing' = standing + ( ( score - standing ) / 10 )


That's not how it works. We have definitive confirmation that it's a average (possibly weighted) of the last X scores, not a running adjustment to a single number. I can't find it right now, but search around a bit, try the wiki first.

Yes, that's what the wiki says,
http://yppedia.puzzlepirates.com/Standing

But there is no ringer statement supporting that, it's a player interpretation; and I have seen arguments why it must work as a continuing adjusted score.

In any case, the sum with i = 1 to 100 of (ith oldest score / 100) has similar potential rounding errors.

Maybe my post should've been the one word 'rounding'.
----------------------------------------
~< The Dread Pirate Chris >~
[Jul 28, 2007 6:07:05 PM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
Shuranthae

Member's Avatar


Joined: Apr 1, 2004
Posts: 9758
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

Well to be frank, score inflation just doesn't make sense if it's not based on a continuing score. I mean it does upto a certain degree but once everyone hits the right level of where their skill should be, things should even out and remain pretty consistent. When a pirate's total history of "X games" gets filled with what they regularly score then it should even out and be consistent. Look at Midnight though, we are still getting score inflation. Sailing and Carp are noticeably harder than I remember it being about four months ago.
[Jul 28, 2007 6:24:48 PM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
Lizthegrey
Developer (retired)
Member's Avatar


Joined: Jul 22, 2004
Posts: 6912
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

* battle_nav                
tgt% act% users total rating
50 50 979 1951 2084
65 65 1270 1951 2224
75 75 1464 1951 2353
85 85 1659 1951 2571
90 90 1757 1951 2713
95 95 1856 1951 2912
99 100 1951 1951 3001


Just quickly ran some numbers to see what the ratings generator was thinking. It looks like in theory, there must be be 19 or fewer ultimates in order for the system to work. However, they ended up being too bunched for some reason, and therefore the threshold was set above the highest score.

mysql> select RATING, count(*) from RATINGS
-> left join USERS on USERS.USER_ID = RATINGS.USER_ID
-> where TYPE_ID = 29 and last_update > now() - interval 10 day
-> and TOKENS & 4 = 0
-> group by RATING order by RATING desc limit 10;
+--------+----------+
| RATING | count(*) |
+--------+----------+
| 3000 | 25 |
| 2999 | 3 |
| 2998 | 1 |
| 2996 | 1 |
| 2995 | 1 |
| 2994 | 2 |
| 2992 | 3 |
| 2991 | 5 |
| 2989 | 2 |
| 2988 | 3 |
+--------+----------+
10 rows in set (0.78 sec)


So... there's your answer.
----------------------------------------
[Edit 4 times, last edit by Lizthegrey at Jul 30, 2007 1:12:52 AM]
[Jul 30, 2007 1:02:47 AM] Show Printable Version of Post        Send Private Message    https://plus.google.com/106912596786226524817/posts [Link]  Go to top 
Shuranthae

Member's Avatar


Joined: Apr 1, 2004
Posts: 9758
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

Well that's certainly enlightening. It also looks freaking crazy. So there really is a max score for ratings? That seems so... not right. I mean, principally not right. Isn't it much better for the design to have a soft ceiling that people slowly reach toward when time equals infinity over a hard cap?

Edit: Wait, I think I see how score inflation is happening now.

I had assumed that the raw scores had no fixed hard ceiling so that scores could rise up to infinity if need be. Because I thought scores could go as high as they can, it wouldn't matter whether or not the raw numbers were relatively or not. However since raw number scores have a ceiling (3000?), that the raw numbers aren't relative is a huge problem.

Basically, at some point in time if this remains the same, it will not only be impossible to get Ultimate in all the puzzles but it will actually be impossible to get Incred or Sparkly.
----------------------------------------
[Edit 1 times, last edit by Shuranthae at Jul 30, 2007 1:52:33 AM]
[Jul 30, 2007 1:47:15 AM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
Rick9109

Member's Avatar


Joined: Sep 19, 2003
Posts: 6716
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

This thread took a turn for the fascinating, and pretty much answered a ton of questions I've always had. Thanks Liz ( =
----------------------------------------
Rome
Pirates of the Damned, Crimson Tide.
I don't care what it did to them, the game's been good to me.
----------------------------------------
[Edit 1 times, last edit by Rick9109 at Jul 30, 2007 4:48:21 AM]
[Jul 30, 2007 4:41:24 AM] Show Printable Version of Post        Send Private Message    rick9109    cactusrome [Link]  Go to top 
Dominator137

Member's Avatar


Joined: Dec 19, 2005
Posts: 1313
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

 
I have to imagine the scores for a Max-0 are fairly close. Time is a factor, it's true, but time is not analog - time would have to be based on turns. Assuming the top bnavers all max-0 brigands in the same number of turns when averaged out is not impossible. I don't know how many active bnavers there are on Midnight, but it could be as simple a matter of 15 people having the same score locking up oceanwide ratings if there are only 1500 active bnavers. If all the top bnavers are really that close, it could vary depending on who has a good day, but it could just rotate in and out, and never unlock itself


Being right FTW?

Anyway, your report is based on a curve, correct? So, is it possible that the 3000 is just how they've performed in relation to the curve? If that was the case, the "hard ceiling" only exists while the curve isn't changing. Assuming that there is more or less a soft cap on how fast you can reliably max and grapple a brigand, it makes sense that there's this congestion, and it also makes sense how we've reached maximum score in relation to the curve.

Of course, I don't know. Just theory.
----------------------------------------
Viridian - Domzx, Senior Officer of The Devil's Own
Member of Regnum Irae

Crooktooth wrote: 
The forum needs a whack-a-mole setting.

[Jul 30, 2007 7:10:54 AM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
sashamorning

Member's Avatar


Joined: Jan 19, 2006
Posts: 3634
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

I agree with Rome. This really is a fascinating conversation.

I have one basic problem, though. Due to the hard ceiling (which seems to have happened because enough players understand exactly what is necessary in order to max out their score in each battle), essentially what the system is saying is that there are no Ultimates because there are too many players that are perfect? With all due respect, that seems backwards. I understand the reason behind the hard cap, but it seems like if you have a max score, you should be seeing the max standing.

I agree with Shur... I expected that the curve would arc up to infinity, although I understand the way it is. If the program looked at the last 100 battles, and 30 was the maximum score for that puzzle, then it's possible that ppl would earn 3000 points and not be able to earn more, simply because once you've done everything that the program expects you to do to meet all the score requirements you can't do much else.

The problem with bnav in context is that it's such a relatively rare puzzle with a long duration. In other words, someone who bilges might have 4-5 more score reports than someone who strictly bnavs. As a result, this kind of congestion is much more likely to happen.

Is there some sort of bnav-related fix that could be coded in, to distinguish the best from the best? There could still be some sort of sliding scale to distinguish those at different levels (like time to grapple, total shots landed, etc. In fact, that'd be interesting... shoot more to overmax, or grapple faster?). Even if the requirements were actually known (I don't expect that), such a system would likely yield SOME sort of delineation between players, at least more than there is now.

That, or give everyone with max score the Ultimate. It's not me, so I'm not campaigning for myself, but if someone has a perfect score, they've done all that the puzzle has requested of them.

(In perspective, a couple of years ago a friend of mine took one of the major grad standardized tests. She scored a 92%ile on one part... but had a perfect score. Her score included a notation stating that although she was in the 92%ile, it was the maximum score... so this does happen from time to time. Related to this, she also got credit for having earned perfect marks, so the %ile wasn't the only thing measuring her success.)

~Sasha
----------------------------------------
Sashamorning - Fiery-haired, fiery-tempered, 100% me.
~~~~~~~~~~~~
Shakespeare wrote: 
This above all: to thine own self be true,
And it must follow, as the night the day,
Thou canst not then be false to any man.

[Jul 30, 2007 9:33:13 AM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
mads0001

Member's Avatar


Joined: Jul 22, 2006
Posts: 6594
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

 


Is there some sort of bnav-related fix that could be coded in, to distinguish the best from the best?

~Sasha


three letters:

P V P

Make pvp the principle factor involved in determining bnav rating. Or a completely seperate standing for pvp. That way, whilst everyone is hammering the AI, at least we'd be able to see who's best at hammering other players...
----------------------------------------
rachaelj wrote: 

"your standing in patronising has gone up and is now respected in your archipelago"
spelling corrected by Mads(tm) :P
Avatar by Funny.
[Jul 30, 2007 9:48:31 AM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
sashamorning

Member's Avatar


Joined: Jan 19, 2006
Posts: 3634
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

 
three letters:

P V P

Make pvp the principle factor involved in determining bnav rating. Or a completely seperate standing for pvp. That way, whilst everyone is hammering the AI, at least we'd be able to see who's best at hammering other players...

GREAT point. But don't make it a separate standing. We'll still have ppl moaning about the current problem. Maybe just factor in PvP scores as worth an extra 25% or something.

~Sasha
----------------------------------------
Sashamorning - Fiery-haired, fiery-tempered, 100% me.
~~~~~~~~~~~~
Shakespeare wrote: 
This above all: to thine own self be true,
And it must follow, as the night the day,
Thou canst not then be false to any man.

[Jul 30, 2007 9:54:40 AM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
SilveRansom

Member's Avatar


Joined: Apr 16, 2006
Posts: 3868
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

 

(In perspective, a couple of years ago a friend of mine took one of the major grad standardized tests. She scored a 92%ile on one part... but had a perfect score. Her score included a notation stating that although she was in the 92%ile, it was the maximum score... so this does happen from time to time. Related to this, she also got credit for having earned perfect marks, so the %ile wasn't the only thing measuring her success.)

~Sasha


/e pokes his abnormally large nose in again and grumbles something about the utter inanity of grading on a curve, since it introduces completely stupid errors like this. How in the world is that an "elegant" mathematical solution?
----------------------------------------
Silveransom
makes art like this or this &
Blender Eggs!
[Jul 30, 2007 10:27:07 AM] Show Printable Version of Post        Send Private Message    http://s13.invisionfree.com/Peregrine_Design/ [Link]  Go to top 
Shuranthae

Member's Avatar


Joined: Apr 1, 2004
Posts: 9758
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

Dom wrote: 
So, is it possible that the 3000 is just how they've performed in relation to the curve?
From what I understand of it, this isn't the case. Raw number scores themselves are no harder or easier to go up or down based on the Ocean. It is harder to go up the higher you are simply because when you're in the top 99% then you have a 1% scoring range and everything else makes you go down, but that's not really based on relativity curves so much.

A good analogy would be to think of an archery or firearm shooting range. The target starts ridiculously large (the side of a barn) and continues to get smaller. As the target gets smaller it's "harder" to hit the target based on your skill, but if you're "good enough" it's really no more or less harder for you. There's a limit to how much smaller it can get and once that limit is reached, it doesn't get any smaller. For our game, it seems like it's a raw score of 3000. The rating number determines the size of your target, the higher your number the smaller the target. There is no other interplay between raw scores, ratings, scoring relativity curves, or what have you for this example.

Let's say that what Standing you get in archery/firearm practice is determined by the size of the target relative to the size of everyone else's target. If you have a tiny target and everyone else has a large one, you'll be Ultimate and be scoring Increds/Sparklies for hitting that tiny target, in theory. There comes a point, if you're good enough, where the target no longer continues to get smaller. Because your target gets no smaller (even if you, yourself, get better) and because anyone not on the high end that continues to practice will get better and will get smaller targets, "score inflation" happens. There eventually comes a time when there's too many players who get the smallest target to justify calling anyone Ultimate or "superior to everyone else." The scoring curve becomes harder as an Ocean gets older not because the high end gets better but because the low and mid end get better with the high end having no where to go.

Now let's say that instead of changing the target size that the target size is fixed. Let's instead change distance with no limit to how distant a target is. As you get better, the target gets farther. Every time you hit the target, it gets farther. Every time you miss, it gets closer. In this scenario there is room for the high end to expand but there is a soft ceiling based on a person's skill. There's a human limit to how far we are able to shoot but that limit isn't so much fixed. This "limit" is much less likely to be the same for a large number of people and it does allow for a more natural sense of progression and better limits score inflation.

Though I'm thinking that score inflation is probably the wrong term for what's happening now that we've seen a bit more of how things work. Still a bit disappointed that it works that way though I suppose it's possible that 3000 just happens to be a soft ceiling for Battle Nav (though I doubt it).

Now that I've typed all this I suddenly realize that I may not have even made the point I was originally trying to make. Oh well, I'll come back when I have more time.
[Jul 30, 2007 10:33:09 AM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
sashamorning

Member's Avatar


Joined: Jan 19, 2006
Posts: 3634
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

I agree with everything Shur just said. I suspect the problem is that there appears to be a maximum score attainable in each battle. It'd be nice if either that was not the case, orif the standards for attaining that maximum score were raised. If there's as much of a clump at the top as there is, then it's reasonable to assume that the maximum score isn't high enough, or that it's just too easy to earn that score. We need more differentiation between the different players.

~Sasha, Solid/Able in bnav.
----------------------------------------
Sashamorning - Fiery-haired, fiery-tempered, 100% me.
~~~~~~~~~~~~
Shakespeare wrote: 
This above all: to thine own self be true,
And it must follow, as the night the day,
Thou canst not then be false to any man.

[Jul 30, 2007 10:39:17 AM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
Chavez67

Member's Avatar


Joined: Aug 26, 2003
Posts: 7739
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

 
I suspect the problem is that there appears to be a maximum score attainable in each battle. It'd be nice if either that was not the case, orif the standards for attaining that maximum score were raised.


I have a hard time believing that Midnight has 25 battlenavvers who's "last X fights" were ALL max-0 vs red-route imperials, which would indicate to me that you can get a 3000 score for less than a perfect battle.

Maybe the scoring system does need work. Maybe at the "X number of fights" in the data pool for calculating rank needs to be increased, to have a higher chance of an imperfect score showing up in the pool.


EDIT: (including some info from emails)

someone forum banned for far less than some others who aren't banned have said, wrote: 
I think that a hard cap of 3000 means that they are using the FIDE system, or something similar. That's a 0-3000 scale, with a mean of 1500. The 3000 point cap is kind of weird - some sources say it exists, while some don't, and there shouldn't be anything in the actual formula that causes a 3000 point cap. Unless you put it in artificially. FIDE did, in fact, switch to a logistic model instead of a normal one.

You can plug in your rating and your opponent's rating to get the expected winning percentage. From there, you can also plug the actual result (0 for a loss, 0.5 for a draw, 1 for a win) into another equation to determine a new rating, based on what the result was supposed to be (somewhere between 0 and 1). So if I was supposed to have a winning percentage of .33 and I win, my rating goes up a large amount. If I draw, my rating goes up a little. If I lose, my rating goes down a medium amount.

What they have probably done is used fractional results. Chess is only win(1)/lose(0)/draw(.5), but bnav can have more result values than that. If a value of 1 is (max-0, less than 15 turns), and a value of 0 if (0-max, greater than 50 turns), everything else can be in the middle - just multiply a fractional term for length of battle by a fractional term for score. So if a hard cap of 3000 has been put in, you get a case where bnavers never go down because they always do at least as well as expected.

Again, this is only assuming they're using a carousing/pvp system (like the swordfighting rankings), rather than a duty system (like for bilge).


That's speculation, and it doesn't exactly jive with the "trailing 12 month graph" type of stat figuring talked about in the wiki, but it might be another piece in the puzzle to figuring out how all this crap works.
----------------------------------------
 
 
 
I did it.
clarify
It was me.

----------------------------------------
[Edit 1 times, last edit by Chavez67 at Jul 30, 2007 11:25:19 AM]
[Jul 30, 2007 10:50:09 AM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
SilveRansom

Member's Avatar


Joined: Apr 16, 2006
Posts: 3868
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

This is pretty much the same thing I was arguing for regarding the scoring and mechanics of Blacksmithing. The scoring needs to have enough variables (ways for people to excel or fail) to make differentiation between players realistic. Trouble is, that requires a lot of time, tuning and coding. At some point, we're probably running into diminishing returns on the Ringer side, as making super fine grained scoring mechanics takes a lot of time with relatively little payoff.

As a game designer and purist, I want that sort of fine grained control, and as a math geek, scoring on a curve bothers me deeply. But, I can understand it if the Ringers don't have time to deal with it.
----------------------------------------
Silveransom
makes art like this or this &
Blender Eggs!
[Jul 30, 2007 10:56:57 AM] Show Printable Version of Post        Send Private Message    http://s13.invisionfree.com/Peregrine_Design/ [Link]  Go to top 
Shuranthae

Member's Avatar


Joined: Apr 1, 2004
Posts: 9758
Status: Offline
Re: Fix for lack of Ultimates. Reply to this Post
Reply with Quote

Chavez, you're assuming that they rate BNav in ways we're not sure they rate it. Quite honestly, I'm starting to believe that it is nothing more than the rating of the bots you fight and your experience. I'm honestly starting to doubt it even matters how well you perform against those bots. At least, in a direct way. Indirectly, if you keep losing you'll spawn weaker bots which then affects your BNav rating.
[Jul 30, 2007 1:57:27 PM] Show Printable Version of Post        Send Private Message [Link]  Go to top 
Posts: 106   Pages: 4   [ First Page | 1 2 3 4 | Next Page | Last Page]
[Show Printable Version of Thread] [Post new Thread]

Puzzle Pirates™ © 2001-2020 Grey Havens, LLC All Rights Reserved.   Terms · Privacy · Affiliates