Renaming "stealth score" to "stealth penalty"

[darksilence]

Any thoughts about this? It's really not a score, but a penalty. It would be much nicer (and less confusing for new players) to see at the end:

Stealth penalty: 0

Maybe "penalty" is not exactly the right word though, but it seems that something like that would fit much better with what that number actually means.

Thoughts?

[Springheel]

Penalties are applied to something.   I don't see how that fits.

[darksilence]

8 hours ago, Springheel said:

Penalties are applied to something.   I don't see how that fits.

Hey, first of all thanks for all of your work! I'm having a blast with The Dark Mod and with learning DarkRadiant (I'm truly enjoying your Speed Build challenge by the way).

I agree that penalty might not be the most appropriate word.

What I meant was that the current stealth score is something that is supposed to be minimized. The breakdown of the score is actually a set of penalty points applied to different situations in which the AI was alerted, so this score is more like an "AI alerted points" (again, not the perfect words, but it's along those lines) than an actual stealth score.

It could be converted into a score if it was presented over a total. For example, 0 could be converted to 100/100, a 1 to 99/100 and so on. That way, when a user sees:

Stealth score: 97/100

they immediately understand it wasn't perfect but it was close. When they see:

Stealth score: 3 

the immediate reaction is confusion and that they played it terribly, especially the first time they see this, but again, if it was more like:

Stealth [insert word referencing something negative that needs to be minimized]: 3 

users could better understand they did minimize what was supposed to be minimized.

Anyways, just my two cents. Thanks for reading.

[Jedi_Wannabe]

I always just thought of it like golf, the lower the better.

[demagogue]

The score was originally a positive score from 0-10, where 10 was a perfect run, and it went down from there. And it turned out that so many games were 0 that people just assumed it was broken, which made it kind of useless. (It also hid the fact that busts were getting overcounted, which is why it was almost always 0 and we didn't even realize it, much less fix it, until we switched to an additive score.) Then we debated increasing it to 0-100 (where 100 is a perfect run), but the thing you have to keep in mind there is "80" will mean something very different for a 10 minute contest sized mission (where it's a really bad score) vs. a 5 hour sprawling mission (where it's a great score).

So to avoid both of those issues I had the idea to make the score positive, so it can go up indefinitely. No matter how bad you are, you always get a meaningful number. And like a golf game there's a "par value" so bigger scores are okay for bigger missions but bad for small missions. So "20" just looks (accurately) like 20 in busts instead of "80" looking (inaccurately) like "80%" stealthy.

When we did that, though, we kept the name "Stealth Score" because it'd already been around for like a year or more at that point. And now it's been around for like 11 years. I understand the logic of the point. "Score" sounds like a higher value is good when it's worse. But the name has a kind of inertia at this point from being around for so long too, and most people understand what it means easily enough.

[nbohr1more]

Heh.

I do kinda agree with the point here.

It should be "Conspicuous Score" or "Clamour Score" etc. AKA, whatever is the antonym of stealth.

[Oktokolo]

I would call it clumsiness score. Bigger is more clumsy.

[stgatilov]

I think it is easy to come up with other words.

For instance, "Stealth flaws" or "Stealth flawness".

[Oktokolo]

How about calling it Flub score?
It just rolls off the tongue, would keep the score, and would still be shorter.

[cabalistic]

I think "conspicuity" or "conspicuity level" would be the most fitting term (as far as I can judge that as a non-native speaker )

[peter_spy]

Idea: it could actually be more fun if the system was somewhat modifiable by mappers, or maybe rewritten entirely. Like having a number of points set by mapper (e.g. 100 or 1000) where detections would deduct from the point total. Although I don't like the fact that even suspicion lvl 2 counts towards that penalty. Searches I can agree with.

Or, maybe focus on positive reinforcement, like in latest Hitman games. Get points for beating a level faster, getting more loot, getting special loot items, performing special actions and finding more secrets. That would encourage replayability, give more stuff to do for more hardcore / competitive players or even speedrunners.

[demagogue]

The score matches the rules for Ghosting, where they know something is there, which fits with the barks for level 2 but not for level 1. If you want to have a rule where just searches and busts count, then you can see that on the front page just by having zero searches and busts.

If you're talking about modifying it, note that it took us weeks of work to get the counting right because of the cascade problem. It's never "one bust" or "search trigger", it's a cascade of like 20 of them in a brief period that looks like "one" to the player.

But what a mapper could do is keep all of that system and just change the actual values of the points for each level though, like make them subtractive instead of additive, or make level 2 busts count for zero, etc. There's a way to get mapper scripts to talk to sourcecode, but I don't remember how to do it. The problem then is that it may be counterproductive to have a different system across maps.

[peter_spy]

By the way:

https://youtu.be/K6y9PJipfpk

I forgot about that grading system in MGSV, and that's interesting because it does let you play more aggressively while it also rewards ghosting approach.

[Bienie]

As an avid ghoster I do very much appreciate the stealth system we have in place at the moment. I think the increasing score system is fine, it might be a bit confusing the first time you play but you get the hang of it quickly. It's only on the end screen anyway where you have to consciously click to see the breakdown. I think the name stealth score is fine, because any word that's the opposite of stealth sounds more like an insult to the player than anything else honestly ^^

I do kind vibe with the par system, I almost wish we had a metric for each map somewhere in the menu system (or even just in the endgame screen) that shows for example your stealth score and then what "par" is, as in a number that the average player would expect to see. That's the only thing I think that is missing from our stealth score system atm, you don't really have anything to compare it to, since the maps vary so wildly it can be hard to tell if it's a good score or not. Personally only a 0 satisfies me but I'm probably in the minority on that one.

 

[Tarhiel]

I try to ghost as much as possible, and even though now I do understand how the score works. but I wouldn´t be against making it vice versa - now the smaller number is better, but it would be easier to understand if 100 (percent) would be optimal and lower score means worse rating.

Edited May 24, 2020 by Tarhiel

[demagogue]

One thing we might do is have the game send the server the stealth score tagged to the mission, and then it could compute a histogram that shows on the end game screen (if they're online), i.e., a bar graph of stealth scores for all players of that FM with a vertical line where the player's stealth score is, so they can see where they rate vis-a-vis the majority. That'd be cool to see and not too hard to set up.

[Silverflame3]

I kinda like the rising stealth score. It reminds of the way linux/unix functions work: If it works properly, a return code given is 0, meaning no faults happened, everything went fine.

Any score higher than that is indicative of the number of failures that happened. This stealth score is, in my opinion, a very good way to show how many "goof-ups" you made during the mission, how severe they were, and I see it as a sort of an inverse scoring board of how "bad" your performance was, like those used in sports competitions.

[kingsal]

I'm glad the scores are there , but I do wish the quick read of the score system was a little more clear.

Stars are pretty universal for ratings , but I'm not sure that feels good thematically. Could work if the art looks cool / appropriate.  Maybe  instead of star icons they are like one of the glyphs or something from thief.   So 5/5 stars is like a 0-10 stealth score, ect..

In the context of a larger campaign , something like  "heat" or a bounty would be interesting. The more foul ups you make the more heat you're getting from the cops . Once gain, maybe  hard to reconcile with the medieval setting.

Edited May 30, 2020 by kingsal

[demagogue]

When I mentioned histogram, I was thinking of the system Opus Magnum uses. It may be easier to just show a screenshot.

In our case, the x-axis would be the stealth score, each unit being a 5 or 10 point range.

The y-axis would just be the raw number of players that fell within that range, with the scale normalized to the max value.

And the player's score would be a vertical line.

I'm imagining that as an FM gets larger, the peak in the curve is going to be pushed further higher (to the right), so when the player's line appears, it's immediately apparent if they're doing better or worse than average, and how much better or worse.

 

 

[kano]

I would change it to "space marine factor". As in, the higher the number, the more like a space marine/doom guy (and less like a thief) you are playing the game.

[chakkman]

On 5/30/2020 at 7:28 PM, kingsal said:

I'm glad the scores are there , but I do wish the quick read of the score system was a little more clear.

Stars are pretty universal for ratings , but I'm not sure that feels good thematically. Could work if the art looks cool / appropriate.  Maybe  instead of star icons they are like one of the glyphs or something from thief.   So 5/5 stars is like a 0-10 stealth score, ect..

I like that idea, would be great.

[thebigh]

Apologies for the partial necro but...

You could even expand the system to an "evidence you were there" counter that would include not only the times you were noticed, but other traces of your presence like doors left open that were closed, mysterious moss patches appearing on the floor, or lights you left extinguished. Obviously this wouldn't include loot or items you took, but might include guards freaking out when they notice something has been stolen (which I think is already the case).

[Darkness_Falls]

Now my apologies for the necro, but I just stumbled on this thread.

I have a feeling the origin or impetus for giving a Stealth Score may have been an idea I had to rate how stealthy the player was while playing the mission, and then give them a career label/title (or some measure-of-success label) with a hand-drawn art graphic characterization of what that rating/title represents.

• My inspiration was Pirates! (old Sid Meier video game) where, when your game concludes, it ranks you (on a societal/class/career scale) based on how well you did and shows artwork/assigns a label to you and the societal class you achieved (e.g., "Beggar" or "Pauper" at the low end vs. "Governor" or "King's Advisor" at the high end).

For The Dark Mod, the labels/ranking would be tailored for the stealth/thief experience. For example, maybe something like "Beggar" at the low end vs. "Master Thief" (for lack of better word) at the high end; with custom artwork made for each ranking showing what you/your status, in theory, looks like at each level.

I always liked Pirates!' ranking system and respective artwork, where you never quite knew what your class rating would be until the game concluded. A nice, tangible visualization of the score... and inherent inspiration/competition to play again and do better next time. (I always wanted to get a higher rating in that game to see the other artworks/labels and get as high as I could get; so it was motivating for me.) It was just one more layer of depth to the game to help you feel as though you were part of that game world.

For The Dark Mod, the thought of doing something like Pirates!, where a bunch of hand-drawn character art would be needed, was just too low on the team's priority list to do at the time; and rightfully so. There were much more important things the team needed to focus on. But I believe that the Stealth Score you see in the game was put in place as a simpler, more feasible, and in-the-interim solution; with the Pirates!-like character art/rankings system possibly being a future evolution possibility, if the opportunity ever arose and/or desire ever came.

I still think something like this would be neat for The Dark Mod, but it's a big ask for an artist or artists to crank out all the character class representations that would be needed

If anyone thinks the Pirates! idea would be neat, I'd be really curious to know what your ideas are for thief/stealth 'classes'. What classes/ranking would you want? What would be appropriate for thievery/stealth? With Pirates!, you see via the link I gave that there are 24 classifications for the 1987 version of the game, and 19 for the 2004 version. They have a lot. Similarly, for The Dark Mod, I think having ~20 classifications would be great, if possible, as it would be nice variety so you're not seeing the same 2 or 3 art graphics every time and so there are several rankings between bottom and top.

 

Edited September 13, 2020 by Darkness_Falls

[wesp5]

26 minutes ago, Darkness_Falls said:

I still think something like this would be neat for The Dark Mod, but it's a big ask for an artist or artists to crank out all the character class representations that would be needed

I agree, that the graphical representation might require too much work, but why not add this just in text form? I like the idea and it shouldn't be too difficult to find class names...

[demagogue]

The problem is that FMs are incommensurable. A "10" would be a fantastic score for a massive FM and a terrible score for a tiny contest-sized FM. So it means somebody would have to play through all of them and personally judge the fair rate of its difficulty in a way that's equal across all FMs, and keep doing it for every new one that comes out. That's a bit unreasonable.

What if the author disagrees with the proposed rate? What if the author's idea for the rate is actually quite inaccurate? If it's such a high burden, you might have multiple people doing it, but then how do you make sure they're being consistent (without just adding more people checking)? Or you just don't, and then the ranking can mean much different things for different FMs. But I think most of all, it's just something that interest will ebb and flow, so work on it may be inconsistent.

All of that said, one might add a cheap normalization system like "Steal Score / number of AI", or probably it'd need to be a little more involved than that, "X - Xmin / Xmax - Xmin", where max and min are ... uh, a function of number of AI... I'd have to think through that. Or have a difficulty coefficient, like 1-3 AI = .1; 4-7 AI = .2; etc... Maybe one could add another modifier based on AI-mutual-proximity (AI next to one another are much more difficult than a large FM with isolated individual AI). Etc. Another problem is that difficulty probably doesn't scale linearly either. But it might be hard to figure out the curve, and how would you measure it anyway? Poll players, rate it by the % of FM time they're "in danger", etc, etc.

It'd take some careful thought, but at least it'd be a universal system that might be somewhat meaningful. But because it'd still be prone to error (even serious error), I think it'd be better to put it in parentheses or in some context to cue the player that the rating may be absurdly wrong. But then players don't understand it and just think it's buggy and we get lots of complaints why their ranking is so ridiculously inaccurate in a given FM and people are going to tell you how stupid and broken the system is, etc... So it may take a lot of iterations to work out the kinks.

Well, if someone wants to try it, they should definitely get a system up and running we could test and work out the kinks over a lot of FMs long before it's going to players. It's hard to judge if it'd be a good system for our game without having a working model we can consider concretely, as opposed to abstract speculation.