The final score - are you happy with how it's worked out?

[sylph]

Since Steam is likely to feature a community hiscore, I just wanted to open a discussion about the score, and whether the community thinks that it accurately reflects a game's achievements?

Here's a rough guideline to some rough 'max' values of scores. Note that these aren't always the absolute maximums, they're just close to the maximum of what has actually been achieved in the community. Some of them might be quite far out (you could certainly get more than 700 knowledge of every spell, for example), but at least the order of magnitude is right...


Category Points My Last Typical Max
Skills skill scores/20 105 125
Spells 50+1/5th of memory per spell 190 3750
piety piety-200, max 10mil 100,000 10,000,000
levelups levelupsx50 2,200 2,450
experience xp/3.5 4,750,000 100,000,000?
exploring all danger levels x30 50,000 76,500
identification IDed items x100 28,000 35,000
intrinsics various, 200-10,000 17,000 35,000?
cute puppy 2k 2,000 2,000
Blup 20k 20,000 20,000
attribute gains 200 per gain (131 average, 26k) 26,000 100,000
Dwarf quests 5k each 25,000 25,000
thieves guild 5-10k (member, head) 10,000 10,000
cat lord 1k 0 1,000
khelavaster 250,000 0 250,000
girdle 2mil 0 2,000,000
hp<300 1k for every hitpoint under 300 0 299,000
atheist 20k for no praying 0 20,000
champion 10-20k (fallen or not) 20,000 20,000
corruptions -20,000 each -200,000 0

closed gate x2
left chain x2 (only if gate is closed)
victory speed 1000 per move under 80k 15,000,000 75,000,000

Despite what the guidebook suggests, it seems that the score is multiplied *before* the speed bonus is added. That means that 75million is about the best score you can get for speedrunning alone, but the 15million for experience and piety will be multiplied up to 60million for finishing the game, rivalling a speedrun!


Have there been any recent changes to the score calculation (ultra endings, corruption, anything else?)?
I saw the following link:http://archive.today/73yTK but it's a bit vague, and I'm assuming the table above is still somewhat accurate...

The most notable point for me is that I think a player's inventory should be auto-sacrificed to add the the piety score, which itself should probably be limited to 5mil instead of it's current astronomical value. I also think the score for exploring, spells, and particularly skills is a little low...


I'd like to try to open a discussion by presenting my take on every section of the final score, but before I do I'd love to see some other opinions on the matter... What do you think of the way the score is worked out? Is it currently in any need of change? Does it represent the difficulty of a given game for typical players?

[_Ln_]

p16:

239202980 Empress (blasph). L50 gray elven wizard (F). 362041022 xps. 223394 turns. Ascended on 3/9/2014.
She managed to become the ultimate ChAoS gOdDeSs in the palace of Andor Drakon.


58767208 Fyrr (ln). L50 dark elven archer (F). 106177088 xps. 90508 turns. Died on 11/16/2013.
She managed to become a ChAoS gOdDeSs in the palace of Andor Drakon.

p21-22:

133016220 Nara (ln). L50 dark elven weaponsmith (F). 29605423 xps. 116992 turns. Ascended on 5/24/2014.
She managed to become the ultimate ChAoS gOdDeSs in the palace of Andor Drakon.


12018591 Elf (asdf). L50 gray elven wizard (M). 26084202 xps. 131829 turns. Ascended on 4/23/2014.
He transformed into an avatar of Order in the palace of Andor Drakon.



62371184 Sidar (jouni). L21 trollish barbarian (M). 4380805 xps. 32336 turns. Won on 7/1/2014.
He saved the world with his brave efforts and became a great ruler while saving himself but once.


Basically, UCGs now trump everything else. I couldn't find Balance Avatars on the servers, so I don't know.

As per topic question, I'd cut out the piety part entirely, upgrade Order and Balance Avatars, RoTMC obtaining score to at least 20-30k.

P.S. Your max score for experience seems off by quite a lot. xp caps on something like 999 999 999 for ~100 mil potential max.

[Blasphemous]

The biggest problem I see with scoring is that Ultra Endings for Order and Balance are poorly rewarded versus normal gate closers. It should be the other way around.
It's probably because of the 2x score for exiting Drakalor Chain after closing the gate. Exiting DC should give some small fixed amount instead of being a multiplier.
Another way would be to multiply score for L and N ultras much like it (probably) is with C ultra.
I don't really understand why UCG is rewarded with such a high score compared to lawful and neutral ultras.

In all cases you need to obtain the trident, get two parts of the trinity and switch alignment, which in the lawful case is a C -> L+ switch, after completing the trinity quests. Not an easy feat I think.
UCG players only need to go from C to N to get the trident and then back to C- to crown - it's much easier to commit a lot of chaotic acts for extreme chaotic alignment than lawful acts to reach L+.
As for the "extremely corrupted" state, it's really easy once D50 is cleared. W5 for a few minutes and voila, or just dip the crown/medal of chaos in any number of useless potions you looted from all the monsters there and drink the PoRC.

Let's be honest - anything you do alignment-wise before char lvl 20 and before the tower is subject to change, bar special cases where paladins are crowned early N= with Justifier.
In that last case, you're not going for ultra anyway so your alignment remains N for the rest of the game.
Chars that need to switch alignment to do ratling and crone quests face a different set of problems because of chaotic requirement, especially lawfuls.
You will lose piety after alignment switch and thus prayers with the lawful deity, which may be quite a setback depending on how prepared you are.
Chaotic chars need at least neutral alignment to get the trident after going through scintillating cave so this is a late game event. Not an easy task to get some semblance of neutrality for the trident at that char level.
This is why I think lawful and chaotic ultras should reward the player with similar score, with UCG giving a bit more due to slightly higher difficulty overall.

EDIT: One more thought: players should be severely penalized for excessively high turncount.
I'm thinking divide the final score by 2 for every 50k turns above 100k turn.
Otherwise grindfest and slow archmage buildup will always win over speed and improvisation as you go.
Empress, that 223k archmage UCG of mine that _Ln_ mentioned, was extremely scummed using all methods possible at that time, with 100+ spare wishes after killing Andor.
That was #1 on the score list and unbeatable at that time.

[Soirana]

Since Steam is likely to feature a community hiscore,

I'd like to try to open a discussion by presenting my take on every section of the final score, but before I do I'd love to see some other opinions on the matter... What do you think of the way the score is worked out?

Assuming players gonna play for highscore and other will get frustrated about highscore inflation... piety has simply either to go away or have limit at like 1% of current limit [otherwise casino gold+altar is easy 10mil], and xp contribution has to have upper limit [don't know - maybe a mil or two] otherwise grindfest for xp is going to trump anything else.
Probably need limit on other stuff like spell memory.

Personally I would say crossing chaos gate should count as exiting chain for score purposes.

Rest not so sure, I mean people will find a way to gamble system anyway so why to bother.

[magpie]

I don't really care.

I don't have Steam and won't ever get it, and I don't play on a server (my latency is abominable, especially with my current location).

And as Soirana said, there'll always be people who get way too far into the metagame of maximising their score by any means necessary rather than having fun with the game itself.

So I don't worry too much about the score - it's nice to beat a previous best, but as for what other people are doing - eh, good for them, but it doesn't affect me.

Let me clarify - I see the score as a way of rating my own games compared to past games I've played, rather than a way to compare myself to others. I'm perfectly willing to concede there's an awful lot of people out there who are better than me at ADoM. :)

If I wanted to compete with others in a game, I wouldn't be playing ADoM.

[JellySlayer]

I don't think that, in a game as complex as ADOM, including a score really makes any sense at all, TBH. In my wins, I pretty much exclusively judge how well I played based on turncount, modified somewhat by class. And whether or not I was doomed, of course. For losses, I usually just go by level/approximate progress.

[Soirana]

I don't think that, in a game as complex as ADOM, including a score really makes any sense at all, TBH. In my wins, I pretty much exclusively judge how well I played based on turncount, modified somewhat by class. And whether or not I was doomed, of course. For losses, I usually just go by level/approximate progress.

Technically there is no reason 'global highscore' could not be fewest turns taken divided into separate subscore lists by class...

Well, I guess that would be promoting certain style of play, but at least it would instantly solve lots of problems.

[Harwin]

Technically there is no reason 'global highscore' could not be fewest turns taken divided into separate subscore lists by class...

Well, I guess that would be promoting certain style of play, but at least it would instantly solve lots of problems.

It certainly makes a very simple, easy to understand high score goal.
I'd separate it by ending as well(at least Ultra vs. OCG vs. Normal). So you aren't competing against an ending with a wildly different expected turncount.

However, it would be nice to reward people (score-wise) for certain trickier things like rescuing puppy. (Maybe those count as equivalent to a certain number of turns)
It also doesn't work for non-winning games.

[YourMum]

Like magpie I've never been interested in the score and judge myself on other things. I think you should get points for completing the different quests even if it's just small amounts. The points you get currently are bizarre and don't seem to relate to how difficult or optional they are.


cute puppy 2k
Blup 20k
Dwarf quests 5k each
thieves guild 5-10k (member, head)
cat lord 1k
khelavaster 250,000

You get 20k for making Blup happy (piece of cake once you have swimming and teleportation) but only 1k for not killing any damn cats?! Yeah the ring is already a decent reward but it can be fairly challenging and annoying (and yes, purely optional but that doesn't matter). You get 5k for killing Yergius once you're in the guild (that is how you become head, isn't it? Not done it in years).

[Harwin]

Cat Lord, Blup, and Khelavaster all provide substantial in-game rewards for their quests.
Puppy doesn't - I feel like it should be worth a lot more points. I guess one problem is that it makes non-complete games that rescue the puppy get much higher scores since it would be a substantial part of your score if it were say, 50k or more, but maybe it could give you bonus points if you win the game.

[boat]

No, I don't think the current high score accurately represents the players achievements. There almost needs to be a calculation matrix to make sure that quests and endings are valued appropriately in their own categories.

I'm from the camp that doesn't care about high score. I think the server has opportunity to provide excellent metrics on quests/race and class combos/ etc and ultimately balancing. Which is good or bad depending how you look at it.

This is a little off the topic of the thread but I suggest it stays because it fuels competition and keeps players around. It does lead to exploiting though as mentioned.

[gut]

I like very little about the scoring system as it is. It rewards things I don't want rewarded, like piety from casino gold, and grinding in general (especially for xp) and doesn't reward things like
challenges and ultras.
For the score to reward the things I think are most fitting, the system would probably be more complex than it is right now. How to reward games like blinging, shield throwing, xl 1 victories, headhunter, or blind monk challenges? One would unavoidably have for these victories a high turn count, mediocre xp, moderate piety, etc... thus always be buried at the middle (if not bottom) of score charts, but at the same time ranked highly in the minds of our fellow adom nerds (which is the true measure of adom greatness, after all). Maybe an analysis of extremes for PC info? This is already the case for turncount, as I believe for every turn < 80K you start getting points.

How about awarding points for game winners under these circumstances:
If total monster kills < x you start getting points.
If total of (weapons skills + magic effectivity points) < x you start getting points. EDIT: this wouldn't work for mindcrafters, I guess
If artifacts genned < x you start getting points.
Maybe if > x you start getting points too?

[gut]

Wait! I left out the most important thing. Corruption points should definitely count as score benefits, not detriments.

[Carter]

would consider having a multiplier for class combos.

gnome wizard = 0.9 vs dark elf merchant at 1.4 or some such. Just to even things up a bit.

[shockeroo]

I agree that current scoring wants to be changed, but I'm not sure how.

I also think there should be seperate high score tables for each of the different endings, and another for losers.

[sylph]

This feedback is more turn-centric than I thought it would be! That makes me smile.

My take on scoring is the following:
I think score should be turncount. Lower turncount, higher score. This is the least scummable, least-grindy, and most relevant metric to a player's skill in the game. It encourages ingenious play, rewards risk, and gives a player a very clear goal to work towards throughout the game, while having little reliance on chance, gaming the system, or simply advantages from knowing the scoring system.
Turncount is already (and has been for a long time) pretty much the metric that top ADOMers use to measure themselves, and trying to get a low turncount refreshes the game, rather than stagnating it the way piety or experience grinding does.

Challenges
I'd like it if challenges were measured with something TB has already announced: Achievements, with a 'challenge' leaderboard simply listing a player name, and how many achievements they have obtained (read below about difficulty). Unlocking achievements (such as 'win the game in under 4 hours playing time', 'reach level 50', 'get a win with a bard', 'score under 70,000 turns', 'become a god' etc) should give the player titles or some other kind of persistent acknowledgement of their accomplishments (just a trophy screen would be fine, but I'll go into titles below). It should be used to keep track of challenge games, and there needn't be a single achievement per challenge.
For example, another favourite roguelike of mine has many lines of achievements, one of which is the 'strongman' line, with an 'easy' achievement for playing a game using only melee attacks, a medium one for winning with only knives, and a 'hard' one for winning with only fists.
Achievements in ADOM could include all the 'ironman' modes, including lithium, eternium, brimstone etc. It could contain 'further' challenges, such as doing ironman using a thief or doing the bug temple with a barbarian, doing the maze without magic mapping etc. It could include combinations of challenges (level 1 challenge + ultra ending), or just simple departures from usual play (reach level 50 without entering the CoC etc).
The important part here, though, is to Categorize these achievements into different 'tiers', so a player like gut can ignore all the lesser challenges and jump straight in at the higher end.

Goals\Rewards:
We could give the player goals to reach for. For example, let a player's classes be changed with titles like 'berserker lord' or 'assassin prince' if they hit certain goals (no ingame changes other than aesthetics, of course)... It could be like: for the title of 'knight', a player must collect 10 achievements, for the title of 'master' you need 8 iron achievements and 2 mithril, for the title of 'lord', you need 5 mithril and 2 adamantium, for the title of 'king', you need 8 adamantium and 3 eternium, for the title of 'emperor', you need an artifact achievement' etc.
The interesting part of 'goals' to reach for the player is that the player still has choice, and is never forced into playing the game in a way they are not comfortable with. If you need another adamantium achievement to progress, but don't fancy gaining 25 levelups while 'very corrupted', you could instead try for the 'no reading' challenge. Making sure a player never has to collect *all* achievements in a given tier ensures that players always have the option of selecting which kind of game they want to play instead of being forced down one road.
Of course, goals and rewards (whether rewards be titles or a trophy room or whatever) would be entirely optional, and a player that wasn't concerned with gaining a 'knight' title or a trophy page, but wanted to compete with other players, could instead just jump in and start aiming for the eternium-level challenges that get into the highscore.
I think that all well-recognised community challenges should probably be put into the achievement system, particularly included those which haven't yet been completed (brimstone?)

Hiscore tables
I think measuring score based purely on the turncount, and achievements based on the other metrics of success, will give players of all levels goals to work towards and unique challenges to overcome, whatever their skill at the game. It also gives a much better feeling of accomplishment and success than a simple (and badly calculated) hiscore table does.
As for the idea of an online leaderboard, we could limit it to simply 2... One leaderboard for turncount (filterable by class/race), and another leaderboard for achievements.
Having appropriately categorised achievements, and a leaderboard that simply sorts on how many a player has obtained at each tier, (so a player with even a single highest-tier achievement would instantly be above any player that had didn't), is a much better way of measuring a player's best moments than trying to quantify exactly whether finishing with a barbarian in under 40,000 turns without praying is worth more or less score than getting a UCG without using spells. It compares players to players, instead of comparing characters to characters, and it gives us a reason to try new things, play outside our comfort zones, and innovate, while leaving a legacy of every character we have ever done well with, instead of just one.

Feedback/discussion:
I asked for answers, and I got quite a few, which I'd like to talk about!

_Ln_:
Thanks for the feedback about experience. I had a fear that it would be capped so highly. I've adjusted the table above. It's particularly worrying that the experience score *is* multiplied => 4x by the 'ending multiplier', while the turncount isn't.

Soirana:
I actually like the idea of piety being limited to a much smaller number than it's current (ridiculous) max, but allowing a player to 'spill over' that limit with their ending equipment (except for gold). This is kind of what I was working towards in my mention of 'automatically sacrificing a player's entire inventory', for the purposes of score, when the game is completed. You're dead right though, the piety (and to a lesser extent other areas of the score) currently ruin the credibility of it somewhat.

Blasphemous:
I smiled a lot when I saw your suggestion for penalising players for excessive turncount. I guess that's the 'stick' where using turncount as score is the 'carrot'. :)

magpie:
I'm totally with your notion about measuring yourself against your past games, and not caring for being competitive (I'm a lot like that). Do you feel that having achievements and milestones to reach in gaining them would be your kind of thing, in that it helps you track your own progress through the different challenges that the game presents?

JellySlayer:
"I pretty much exclusively judge how well I played based on turncount".
I have a lot of respect for that. :)
Would challenges as I described above be of any interest to you then? Or would you just keep trying to better your turncount if you were playing in an environment like the one I suggested?

Harwin:
"I'd separate it by ending as well(at least Ultra vs. OCG vs. Normal). So you aren't competing against an ending with a wildly different expected turncount."
This is a good point. Maybe we could filter the hiscore list based on race, class, and ending type?
"However, it would be nice to reward people (score-wise) for certain trickier things like rescuing puppy."
That would be a good 'starter' for the lowest tier of achievements!
"It also doesn't work for non-winning games."
Another really good point! Maybe the existing score should be kept only for characters that didn't win at all? I probably think that all non-victories should be below victories in the hiscore table anyway!

YourMum:
"You get 20k for making Blup happy (piece of cake once you have swimming and teleportation) but only 1k for not killing any damn cats?!"
Maybe Thomas didn't want to give out score to players for NOT doing things? ;)

Boat:
"I think the server has opportunity to provide excellent metrics on quests/race and class combos/ etc and ultimately balancing. "
I, on the other hand, don't think 'balancing' (at least in terms of race and class) needs to happen at all. In fact I hope it never does! I consider race/class your 'easy, medium, hard' mode in ADOM.

Gut:
I think measuring based on their most extreme character facet is a really interesting way of measuring player skill. As you said it already happens for turncount.
Since your post was concerned, probably more than any others, with the notion of measuring player accomplishments in the minds of the community, how do you feel about what my suggestions would do for the sense of competition?
In other achievement-based scoring systems, players tend to have signatures displaying how many achievements they have in each category, and in my experience it works remarkably well to help players measure themselves and respect the abilities of one another, but would it work for ADOM?

Carter:
That's not actually too bad an idea (a multiplier for race/class). Trouble is, past a certain point, most classes tend to 'blend into one' in ADOM. By D:30, pretty much the only big difference between races/classes is whether they can cast a lot of spells or not.

[Al-Khwarizmi]

Personally I've never cared much about the score, but that is precisely because I find it quite arbitrary. If the score calculation were more meaningful, it would be interesting to challenge myself to improve the score.

I think these suggestions by Soirana would be steps in the right direction:


Assuming players gonna play for highscore and other will get frustrated about highscore inflation... piety has simply either to go away or have limit at like 1% of current limit [otherwise casino gold+altar is easy 10mil], and xp contribution has to have upper limit [don't know - maybe a mil or two] otherwise grindfest for xp is going to trump anything else.
Probably need limit on other stuff like spell memory.

Personally I would say crossing chaos gate should count as exiting chain for score purposes.

Also, finding orbs should give a noticeable amount of points, as the number of orbs found is probably the best indicator of progress in the plot for players that don't finish the game.

[Harwin]

I like Sylph's point of making the challenge games achievements, and having them separately.

And for those who want to win the game with every class, DOOMED, the way I've seen achievements done like that before is:
[Tier1]Win the game with 1 class while DOOMED.
[Tier2]Win the game with 5 different classes while DOOMED.
[Tier3]Win the game with every class while DOOMED.

And the achievement has a list of every class, with checkmarks/highlight/bolding/whatever of the ones you've completed. A ranking that showed achievements completed would be great. Is "acid spit" harder than "bling"? How much harder? You wouldn't have to assign relative scores to them - you could check who had completed them.

[Blasphemous]

I don't mind getting very low scores for excessive turncount, as I love lengthy games, I love exploring and collecting and killing.
I don't think there should be such a high pressure on finishing early, turncount-wise - it encourages skipping optional content and favors speedrunners.
Of course, speedrunners are almost universally the best players in any game (just take a look at speeddemosarchive(dot)com).
This however makes the scoring system a feast for elitists that know every single detail of the game and use that to skip large portions of mentioned game without exploring - something I can hardly imagine that new or moderately advanced players would do.

I'm all in for penalizing grinders but let's not reward pure speed runners that much, otherwise the highscore will have the same people all the time.

[_Ln_]

I like Sylph's point of making the challenge games achievements, and having them separately.

And for those who want to win the game with every class, DOOMED, the way I've seen achievements done like that before is:
[Tier1]Win the game with 1 class while DOOMED.
[Tier2]Win the game with 5 different classes while DOOMED.
[Tier3]Win the game with every class while DOOMED.

And the achievement has a list of every class, with checkmarks/highlight/bolding/whatever of the ones you've completed. A ranking that showed achievements completed would be great. Is "acid spit" harder than "bling"? How much harder? You wouldn't have to assign relative scores to them - you could check who had completed them.

Acid spit and blinging were both completed in 1.1.1 and seriously annoying now:
1) No starvation training means that you need to farm strength potions somewhere.
2) Dual-wielding shields (which is the only way to avoid making a melee hit) brings missile combat to nothingness thanks to dynamic to-hit penalty. However, if there would be an option to install restrictions in-game (such as even if you bump into something with stuff equipped you won't him a monster thanks to your in-game character code of honor or smth), this would make blinging reasonable again (trust me, I tried it in 1.2.0 and it works precisely because every now and again you accidentally hit something).

[Soirana]

otherwise the highscore will have the same people all the time.

That is pretty much guaranteed under any system.

[Al-Khwarizmi]

That's not bad as long as

(1) the highscore holds scores for every player (most of us won't be competing for being the top player, but maybe we can compete for the top 100, top 1000, etc.), and even better,

(2) if we can get regional highscores, monthly highscores, Steam friend highscores, etc.

[psy_wombats]

I've actually been thinking about this a bit lately but couldn't come to a conclusion, but some thoughts:

Now that I've won a bunch, I care a lot more about that score than I did years ago when I started playing, but the way the score system is set up now has some adverse incentives, especially about turncount. In ~20 wins I've only ever done the minotaur lair once, rift once or twice, blue dragon caves once or twice, etc, just because the score I'd gain from exploring those locations is not worth the turns spent clearing them. Same goes for some of the more fun things in ADOM like greater vaults or the ice queen quests: I don't really enjoy choosing between the fun parts of the game and score if I care about score.

I don't really know how to solve this (other than maybe deemphasizing turncount a bit) but I think it might be fun to add score values for some of the optional locations in the game. For example, if it's worth 500,000 points to visit the library and I think I can do that in 500 turns, then maybe it's worth a try. Of course this would require lowering the score-per-turn value from 1000 to something lower and moving the 80k score cap up to maybe 100k or 150k or something so that players that exceed 80k turns still have an incentive not to grind.

[gut]

> 2) Dual-wielding shields (which is the only way to avoid making a melee hit

Gold pieces. Besides, part of the challenge is to keep from bumping into anything. It is not unreasonable to expect players to be perfect :D

[grobblewobble]

I think score should be turncount. Lower turncount, higher score.
That would make sense and I like the idea, but there is a problem with it. Score also needs to make sense for characters that don't win. The score needs to grow progressively for characters that died, as they achieved more and got further into the game. I can't really think of any other way to do that, except with a similar system as we have now.

But yeah, the xp part and the piety part badly reward grinding and need to go. And UCG shouldn't score so much higher than any other ending, I think that's a bug.

[Harwin]

That would make sense and I like the idea, but there is a problem with it. Score also needs to make sense for characters that don't win. The score needs to grow progressively for characters that died, as they achieved more and got further into the game. I can't really think of any other way to do that, except with a similar system as we have now.

But yeah, the xp part and the piety part badly reward grinding and need to go. And UCG shouldn't score so much higher than any other ending, I think that's a bug.

Winning scores don't have to be on the same scoreboard as non-winning scores.