Streakscumming/Winratescumming


If it doesn't fit anywhere else, it belongs here. Also, come here if you just need to get hammered.

User avatar

Tomb Titivator

Posts: 802

Joined: Sunday, 30th March 2014, 21:06

Post Friday, 30th January 2015, 15:51

Streakscumming/Winratescumming

Here are my thoughts on 'winscumming', specifically in terms of streaking and winrate.

Let's suppose that any player of reasonable skill and infinite patience can win a game with careful play if he/she survives past the very early game. Let's also suppose that there is a % chance of unavoidable death in the early game, depending on race and background. Some characters, i.e. MiBe and any troll, have virtually no chance of unavoidable death early game.

A player of reasonable skill could start a new account and steak MiBe HOBe and Trolls for literally hundreds of games because there is little chance of early death.

Now, let's say that the same player decides to create multiple streaking accounts with random backgrounds, and immediately move on to the next one once a streak is broken.
I can imagine that there would be a normal distribution of streak lengths based on the difficulties of the backgrounds chosen and some luck in the early dungeon. Only the longest ones would garner any attention.

How does this affect anyone else who plays crawl? Not in the slightest.
Does anyone actually do this? A few people. Maybe. It seems a lot less trouble just to concentrate the effort to a single account.
What is the effect if many people were to do this? There would be quite a few accounts with high winrates and many wins.
Should the practice be discouraged? No. If that's how people enjoy playing, good for them.

What does having a long streak mean? Only that the player's combination of early game luck and overall skill were sufficient to achieve a streak of that length.
Comborobin Admin

Vaults Vanquisher

Posts: 431

Joined: Saturday, 9th November 2013, 14:34

Post Friday, 30th January 2015, 16:16

Re: Streakscumming/Winratescumming

It's not clear to me how multiple accounts would change the distribution of streak lengths and how this links to certain backgrounds being more likely to result in a win. Playing 'easier' combinations would tend to result in longer win streaks, on average, but distributing across multiple accounts shouldn't have any impact whatsoever, unless you focus on particular combinations for particular accounts.

Ziggurat Zagger

Posts: 3037

Joined: Sunday, 2nd January 2011, 02:06

Post Friday, 30th January 2015, 16:24

Re: Streakscumming/Winratescumming

Winning this game even once, ever, is actually pretty impressive when you get right down to it, no matter how many attempts it took you to get there. Any streak of reasonable length is actually pretty remarkable, and the degree to which a streak is impressive easily eclipses any possible benefit the player could have gotten from accounting shenanigans.

I think people who have yet to win even one game should probably stop trying to crap all over the achievements of others. They're mostly just splattering themselves, and it certainly isn't getting on any of the streak-leaders.

For this message the author KoboldLord has received thanks: 2
Arrhythmia, nago
User avatar

Tomb Titivator

Posts: 802

Joined: Sunday, 30th March 2014, 21:06

Post Friday, 30th January 2015, 16:31

Re: Streakscumming/Winratescumming

Jarlyk wrote:It's not clear to me how multiple accounts would change the distribution of streak lengths and how this links to certain backgrounds being more likely to result in a win. Playing 'easier' combinations would tend to result in longer win streaks, on average, but distributing across multiple accounts shouldn't have any impact whatsoever, unless you focus on particular combinations for particular accounts.


True, multiple accounts would not change the distribution of streaks, but each account would show a single point on the player's streaking bell-curve. So that you wouldn't see all the shorter streaks on the path to the big one.
Comborobin Admin

Crypt Cleanser

Posts: 723

Joined: Monday, 9th June 2014, 13:39

Post Friday, 30th January 2015, 16:33

Re: Streakscumming/Winratescumming

Okay, xentronium, thanks for starting a new thread. Maybe a mod can move the relevant stuff to the top of this one; I don't know. But since it's here, a little background might help better clarify my position and exactly what I mean by "winscumming."

As I mentioned elsewhere, I was a Diablo II player for many long years. Before that, I was a Moria player for many long years. Pretty much those were it. I'm not even a "gamer" really by that metric. In both of those games, I played "scummy;" that is, in Moria I alternated between real play and savefile backup play. In Diablo, there was no such distinction. They allowed for both kinds of play as a matter of personal preference: hardcore or softcore. I always preferred softcore, and just couldn't stand the idea of losing everything by something as simple as dying. As wheals mentioned, this is a personal preference, and one that Blizzard simply recognized and catered to. But then, Blizzard lost the beat, did a bunch of stuff I thought was unethical, and took a big fat dump on their entire legitimate player base, so I quit playing their game in protest. Then, I found Crawl.

At first, when I read the design philosophy and first encountered the term, savescumming, I was offended. Who were they do debase those who preferred a different style? It put me off quite a bit when I got online and realized "scumming" was attached to all kinds of different behaviors. It was clearly a philosophical underpinning that such behavior is anathema in all its forms. So I continued to play offline, wondering what was wrong with those people.

Here is what I mean by playing offline: I played one character, over and over and over, scumming everything to the max, including mutations (eating chunks, quitting and eating again until good ones showed), zig floors (quitting until getting the loot i wanted), and on. After finally managing to "win" with that character (which was incredibly difficult, to me), I literally could not believe that anyone could win the game without doing that. So I got online, and started playing to win legitimately. I decided to adopt the philosophy of Crawl, and play it as it was intended. I became a convert.

Now, perhaps the worst kind of believer is the convert, because in the fervor of adopting new beliefs and in the process of internalizing new philosophies, one inevitably notices the contradictions that longtime believers may never give a moment's thought to. I almost immediately saw how the practice of manipulating one's winrates was philosophically identical to any other kind of manipulation to improve results, and I first posted my observation here, several moons ago, soon after first joining.

This is not rocket science, but it is also not a judgment of preference for my choice over anyone else's. It is rather an observation about the prejudice inherent in the term itself, in that it allows a double standard to operate unquestioned because the practitioners are the ones defining the game.

You really can't have it both ways. If you are going to sneer and shudder at certain kinds of behaviors, then to not do so universally is hypocritical. In the end, that is the entire problem with rigid philosophies, because there are always exceptions to man-made rules.

I have my wins now to prove it possible, and I continue playing. But comparison of my winrate to anyone else's winrate will never reflect anything about Crawls's difficulty, no matter how many times the argument is raised. Unless winrates cannot be manipulated, they are meaningless. That's all.

For this message the author Aule has received thanks:
Lyrick

Ziggurat Zagger

Posts: 11111

Joined: Friday, 8th February 2013, 12:00

Post Friday, 30th January 2015, 16:46

Re: Streakscumming/Winratescumming

Aule wrote:But comparison of my winrate to anyone else's winrate will never reflect anything about Crawls's difficulty, no matter how many times the argument is raised. Unless winrates cannot be manipulated, they are meaningless. That's all.


Well, it depends. If you see a player with 99 wins in 100 games, does it tell anything about crawl's difficulty? I hope it does.
Also we have "Best score" and "Streaks" sections for every player, are they meaningless to you too?

Vaults Vanquisher

Posts: 431

Joined: Saturday, 9th November 2013, 14:34

Post Friday, 30th January 2015, 17:14

Re: Streakscumming/Winratescumming

Ah, I see what you're suggesting. The tricky thing with statistics is that it really doesn't make sense to compare two mean values in and of themselves when there is randomness at play. If a player splits their play among multiple accounts, there will be some distribution of win rates, with some having higher values than others. However, assuming we want to compare with an equivalent single account, they would also have a fewer number of games per account, which means the estimator of 'true win rate' (in other words, how likely the player would be to win, going forward, assuming they play under the 'same' conditions with the 'same' skill) has a higher random variance.

In practice, of course, this is complicated by that fact that skill level presumably doesn't stay fixed over time (players learn to play better through practice), conditions are not equal (choice of starting combination has an impact on win rate) and the difficulty of the game presumably fluctuates over time as it is updated. These effects violate the assumption of win rate being a metric of independent random events, which means even taking into account random deviation would not make it a strong predictor of future win rate. This isn't to say that it's a useless metric if you for some reason wanted to predict a player's future win rate (say, you're taking bets and want to set the house odds), but the predictive power is likely to be fairly low.

None of this is meant to denigrate the skilled players who can achieve high win rates consistently, but just thought I'd chime in a bit on the statistics aspect of this. If you really wanted to compute the probability of one skilled player being more likely to win in the future than another skilled player, consider it on the same order of difficulty as attempting to place bets on the performance of pro sports players.

Tomb Titivator

Posts: 909

Joined: Thursday, 3rd January 2013, 20:32

Post Friday, 30th January 2015, 17:20

Re: Streakscumming/Winratescumming

Sandman25 wrote:
Aule wrote:But comparison of my winrate to anyone else's winrate will never reflect anything about Crawls's difficulty, no matter how many times the argument is raised. Unless winrates cannot be manipulated, they are meaningless. That's all.


Well, it depends. If you see a player with 99 wins in 100 games, does it tell anything about crawl's difficulty? I hope it does.
Also we have "Best score" and "Streaks" sections for every player, are they meaningless to you too?

Aule: Being able to manipulate something doesn't make it "meaningless" -- unless you can make it literally any value you want, with minimal effort. The kind of "winscummy" behavior you're talking about is not effortless and cannot produce completely arbitrary values, though it can influence those values pretty strongly in some cases.

Sandman: A single data point is unlikely to tell anybody anything about anything.

Anyway, let's just all agree that winrates considered in isolation are not the best metric for Crawl difficulty and leave it at that, hmm?
Wins (Does not include my GrEE^Veh 15-runer...stupid experimental branch)

For this message the author tedric has received thanks:
xentronium

Ziggurat Zagger

Posts: 11111

Joined: Friday, 8th February 2013, 12:00

Post Friday, 30th January 2015, 18:01

Re: Streakscumming/Winratescumming

tedric wrote:Sandman: A single data point is unlikely to tell anybody anything about anything


...unless it's a personal record. Personal records are often shown on Olympics TV broadcasts etc. If you see one person with personal record 100 and another with 80 (more is better), then it's unlikely that the second will win.

Tomb Titivator

Posts: 909

Joined: Thursday, 3rd January 2013, 20:32

Post Friday, 30th January 2015, 18:05

Re: Streakscumming/Winratescumming

That's two data points. And yet it doesn't tell you anything about how hard the sport is.
Wins (Does not include my GrEE^Veh 15-runer...stupid experimental branch)

Ziggurat Zagger

Posts: 6454

Joined: Tuesday, 30th October 2012, 19:06

Post Friday, 30th January 2015, 18:08

Re: Streakscumming/Winratescumming

Sandman25 wrote:
tedric wrote:Sandman: A single data point is unlikely to tell anybody anything about anything


...unless it's a personal record. Personal records are often shown on Olympics TV broadcasts etc. If you see one person with personal record 100 and another with 80 (more is better), then it's unlikely that the second will win.

That's not true, without more information you can't say whether a personal record is indicitive of a skill level.

In your example, what if the "score" is percentage, but person 1 has only played one game ever, and person 2 has played 1000.

Without knowledge of sample size and how it compares and how the score is calculated the two "scores" are meaningless, even as indicators of personal performance.

Without context and knowledge of how the numbers were created, the are meaningless.
Spoiler: show
This high quality signature has been hidden for your protection. To unlock it's secret, send 3 easy payments of $9.99 to me, by way of your nearest theta band or ley line. Complete your transmission by midnight tonight for a special free gift!

Ziggurat Zagger

Posts: 11111

Joined: Friday, 8th February 2013, 12:00

Post Friday, 30th January 2015, 18:09

Re: Streakscumming/Winratescumming

tedric wrote:That's two data points. And yet it doesn't tell you anything about how hard the sport is.


Well, it depends. If you see a player with 99 wins in 100 games , does it tell anything about crawl's difficulty? I hope it does.
Also we have "Best score" and "Streaks" sections for every player, are they meaningless to you too?

Ziggurat Zagger

Posts: 11111

Joined: Friday, 8th February 2013, 12:00

Post Friday, 30th January 2015, 18:13

Re: Streakscumming/Winratescumming

Siegurt wrote:In your example, what if the "score" is percentage, but person 1 has only played one game ever, and person 2 has played 1000.


I doubt I would see a sportsman with one game in Olympics broadcast.
Number of played games is present in player profiles too.
Let me put it another way. Do you think all info in player profile is meaningless and tells nothing about player (i.e. you can never compare 2 players looking at their profiles)?

Ziggurat Zagger

Posts: 6454

Joined: Tuesday, 30th October 2012, 19:06

Post Friday, 30th January 2015, 18:21

Re: Streakscumming/Winratescumming

Also attempting to "winscum" as described by the OP is fairly pointless, yes, you can create new accounts in an attempt to "game" the system. However no-one cares (Even "bragging rights" are pretty much dismissed and uninteresting) so if someone makes new accounts in order to try to game themselves into a high-percentage winning (or large-streak making) record, you know what, no one cares about their "Impressive" record.

If someone is attempting to use winrates to "prove" a point about how easy or hard DCSS is, they're blowing hot air, using statistics to "prove" a point is the laziest and least-meaningful form of argument.

Even in a completely "fair" system where it's difficult or impossible to game the system to produce outlying results, it's always possible to interpret the data to match one's preferred point of view. Statistics, at best, can be used to *correlate* real-life results with logic and well-reasoned arguments, they can never form a solid foundation.

I am under the impression that the vast majority understand this (if not explicitly, then implicitly at least) and it's only a small minority that slavishly attempt to "prove" things using stastics, typically to back up points they can't argue rationally and logically.
Spoiler: show
This high quality signature has been hidden for your protection. To unlock it's secret, send 3 easy payments of $9.99 to me, by way of your nearest theta band or ley line. Complete your transmission by midnight tonight for a special free gift!

For this message the author Siegurt has received thanks:
gammafunk

Ziggurat Zagger

Posts: 6454

Joined: Tuesday, 30th October 2012, 19:06

Post Friday, 30th January 2015, 18:23

Re: Streakscumming/Winratescumming

Sandman25 wrote:
Siegurt wrote:In your example, what if the "score" is percentage, but person 1 has only played one game ever, and person 2 has played 1000.


I doubt I would see a sportsman with one game in Olympics broadcast.
Number of played games is present in player profiles too.
Let me put it another way. Do you think all info in player profile is meaningless and tells nothing about player (i.e. you can never compare 2 players looking at their profiles)?

I didn't claim you could never compare two players by looking at their profiles, only that you had to understand what the numbers meant and how they were arrived at to understand how significant they are and how much weight to give them.
Spoiler: show
This high quality signature has been hidden for your protection. To unlock it's secret, send 3 easy payments of $9.99 to me, by way of your nearest theta band or ley line. Complete your transmission by midnight tonight for a special free gift!

Ziggurat Zagger

Posts: 11111

Joined: Friday, 8th February 2013, 12:00

Post Friday, 30th January 2015, 18:27

Re: Streakscumming/Winratescumming

Siegurt wrote:I didn't claim you could never compare two players by looking at their profiles, only that you had to understand what the numbers meant and how they were arrived at to understand how significant they are and how much weight to give them.


Sorry, that's trivial. We are not talking about idiots here I hope.

Halls Hopper

Posts: 76

Joined: Wednesday, 5th March 2014, 21:07

Post Friday, 30th January 2015, 18:29

Re: Streakscumming/Winratescumming

Aule wrote:You really can't have it both ways. If you are going to sneer and shudder at certain kinds of behaviors, then to not do so universally is hypocritical.

There's a lot of words, so I'll focus on that point since it seems to be the root of contention.

I don't see any sneering or shuddering or whatever toward grinding amongst the dev team. I'd be surprised if it did exist, since a number of them play and have played Diablo, Angband, NetHack, et cetera. But since a game can't please everyone due to conflicting tastes, Crawl development works to make "optimal" play not involve grinding.

Regarding "optimal": I understand that to be what makes it most likely to win a specific game. At least, I think there was a specific type of "optimal" intended when Crawl's philosophy was devised; otherwise, you could hold any type of strange behaviour against the game because it is considered optimal under one distorted definition. And "winscumming"" sounds like you're explicitly throwing away some games, so it's not a good way to win a specific game.

For this message the author basil has received thanks:
Sar
User avatar

Shoals Surfer

Posts: 301

Joined: Friday, 8th November 2013, 16:19

Location: Tel'aran'rhiod

Post Friday, 30th January 2015, 18:32

Re: Streakscumming/Winratescumming

I really want to know how scumming for better streaks works...

For this message the author Tedronai has received thanks:
Sprucery

Ziggurat Zagger

Posts: 6454

Joined: Tuesday, 30th October 2012, 19:06

Post Friday, 30th January 2015, 18:47

Re: Streakscumming/Winratescumming

Sandman25 wrote:
Siegurt wrote:I didn't claim you could never compare two players by looking at their profiles, only that you had to understand what the numbers meant and how they were arrived at to understand how significant they are and how much weight to give them.


Sorry, that's trivial. We are not talking about idiots here I hope.

Well, it may be a trivial point to you, but there's a lot of people out there who whip out a pair of numbers and say "See look this *means* something" with no understanding of where the numbers come from or how they relate or what they, in actual fact, mean, or equally badly, *do* understand what the numbers mean, but use them as a way to argue points with people who don't (or in some cases *can't*) know how the numbers were arrived at or what they mean.

I'm not sure I would label all people who do so "idiots" (Although to be fair there are a lot of idiots in the world) I would say that many people are so wrapped up in a myopic limited point of view that they can't see how anyone else could interpret things differently than they do. They take numbers as gospel because their interpretation backs up the points they believe in. The problem with that is that *even when they are right* it's not the numbers that communicate or mean anything *in and of themselves* it's the underlying meaning that needs to be communicated.
Spoiler: show
This high quality signature has been hidden for your protection. To unlock it's secret, send 3 easy payments of $9.99 to me, by way of your nearest theta band or ley line. Complete your transmission by midnight tonight for a special free gift!

For this message the author Siegurt has received thanks:
Sandman25

Barkeep

Posts: 3890

Joined: Wednesday, 14th August 2013, 23:25

Location: USA

Post Friday, 30th January 2015, 18:49

Re: Streakscumming/Winratescumming

See also these posts split off from poor Bloax's thread about his magnum Opus.

FWIW anyone who thinks that having multiple accounts is "winscumming" or whatever is putting way too much stock in winrates as a statistic, and cares about other people's winrates way too much.

For this message the author and into has received thanks:
Sandman25

Ziggurat Zagger

Posts: 11111

Joined: Friday, 8th February 2013, 12:00

Post Friday, 30th January 2015, 19:00

Re: Streakscumming/Winratescumming

Siegurt wrote:I'm not sure I would label all people who do so "idiots" (Although to be fair there are a lot of idiots in the world) I would say that many people are so wrapped up in a myopic limited point of view that they can't see how anyone else could interpret things differently than they do.


It's first time when I am called idiot (ok, you wrote "I'm not sure I would label") because I expect opponents to be non-idiots and to recognize the obvious. Just kidding, I hope I understand what you mean and I agree.

Tomb Titivator

Posts: 853

Joined: Thursday, 29th August 2013, 18:39

Post Friday, 30th January 2015, 20:34

Re: Streakscumming/Winratescumming

Not sure why there are two threads, but see my post in the other thread. Lifetime winrate is not a meaningful measure of the ability of a player. Longest streak is much closer but still not quite there. Winrate as computed by a moving average with higher weighting towards more recent games is probably the best one. All that lifetime winrate truly measures is how long the player took to learn the game, not how good they are at it, and to me that's not a very interesting thing to talk about.

For this message the author johlstei has received thanks: 3
Brannock, Sandman25, xentronium

Dungeon Master

Posts: 3160

Joined: Sunday, 5th August 2012, 14:52

Post Friday, 30th January 2015, 20:41

Re: Streakscumming/Winratescumming

Aule wrote:Now, perhaps the worst kind of believer is the convert, because in the fervor of adopting new beliefs and in the process of internalizing new philosophies, one inevitably notices the contradictions that longtime believers may never give a moment's thought to.

It sounds like you're saying that you have a clearer perspective because, unlike other posters, you have an outsider's perspective. I hope you don't genuinely believe either that being an outsider makes your perspective clearer or that you're the most "outsider" person in the discussion.

Aule wrote:I almost immediately saw how the practice of manipulating one's winrates was philosophically identical to any other kind of manipulation to improve results, and I first posted my observation here, several moons ago, soon after first joining.

I'm not sure why you bothered to provide that link. That you chose to say the phrase winscumming a while back is not meaningful to this discussion.

You've asserted that manipulating winrates can be done, but you haven't given any support for an argument that it's meaningfully possible. Obviously, it's trivial to create accounts and play one game on each one until you win a game, and then claim you have a winrate of 100%, but equally obviously that's meaningless and not relevant to any discussion at hand. Everyone understands that win ratio is impressive in proportion to the number of games across which it applies, the types of characters played in those games, and the voluntary challenges undertaken in those games. Whenever anyone (or anyone else, at any rate) refers to winrate or streak length, they are assuming they don't have to spell out these caveats because they're obvious.

Aule wrote:But comparison of my winrate to anyone else's winrate will never reflect anything about Crawls's difficulty, no matter how many times the argument is raised. Unless winrates cannot be manipulated, they are meaningless. That's all.

A (purely hypothetical) winrate of 1000 wins over 1000 games with a broad variety of characters would be powerful evidence that crawl is always winnable, and that the player responsible for the account is a player who can choose to win just about every game. There's no way for a player to game that, aside from having root access to the server and generating huge amounts of detailed fake data. You've mentioned offline and wizard mode previously, but neither offline nor wizard mode games appear in these statistics.

For this message the author Lasty has received thanks:
Sandman25

Tomb Titivator

Posts: 853

Joined: Thursday, 29th August 2013, 18:39

Post Friday, 30th January 2015, 20:44

Re: Streakscumming/Winratescumming

Like here, take this exercise:

Player A takes 1000 games to learn, with a 10% winrate during that time. Then they proceed to win 100 games in a row, for a 100-streak. Then, they play 500 more games, winning 98% of them, losses distributed uniformly across that time.

Player B takes 100 games to learn, also with a 10% winrate during that time. Then they proceed to win 110 games in a row, for a 110-streak. Then, they play 500 more games, winning 95% of them, losses distributed uniformly across that time.

Assume a similar distribution of character types between both of them. Both players are playing to win, no one start scums or quits or anything like that. They are playing the same version across all games.

Which player is better? To me, "better" implies a higher likelyhood of winning a given game. That is, even though the player B has a much higher lifetime winrate, and a longer streak, I would say undoubtedly player A is better, since they are more likely to win a given game, at 98% vs 95%. If you disagree, feel free to say why, if you don't, should you not also agree that lifetime winrate ought to be discounted and there should be no such notion as "scumming" it?
User avatar

Barkeep

Posts: 4435

Joined: Tuesday, 11th January 2011, 12:28

Post Friday, 30th January 2015, 20:53

Re: Streakscumming/Winratescumming

Tedronai wrote:I really want to know how scumming for better streaks works...

You play a game. If you win, you play another one, and so on. If at any point you lose, you play another game, until you start winning again. Do it enough, and eventually by random chance you will get a really long streak of wins.

Welp, that's it. I'm headed to Vegas; wish me luck!
I am not a very good player. My mouth is a foul pit of LIES. KNOW THIS.
User avatar

Shoals Surfer

Posts: 301

Joined: Friday, 8th November 2013, 16:19

Location: Tel'aran'rhiod

Post Friday, 30th January 2015, 22:10

Re: Streakscumming/Winratescumming

njvack wrote:You play a game. If you win, you play another one, and so on. If at any point you lose, you play another game, until you start winning again. Do it enough, and eventually by random chance you will get a really long streak of wins.

Damn, it sounds really easy when you put it like that... I feel enlightened! :)

Halls Hopper

Posts: 88

Joined: Monday, 8th July 2013, 08:17

Post Friday, 30th January 2015, 22:15

Re: Streakscumming/Winratescumming

The reasons why players might look down on savescumming or similar kinds of scumming vs. streakscumming strike me as very different.

When players do things like savescum, a lot of times they're inhibiting themselves from becoming better at the game. Sure, when I finally won Nethack when I was a kid, it might have seemed like an accomplishment at face value. But I used savescumming as a crutch to get there. It wasn't until I stopped the practice that I got good enough to win without it, because.... well, without being able to revert to a previous save if things went south, I had to learn how to avoid death.

It's the same sort of impact that using cheat codes can have. Sure, you will be able to win a game, but... it's a crutch and it can prevent you from honing actual skill.

Streakscumming is just number manipulation. Whether or not you win 20 consecutive games with MiBe's does not crib your skill in the same way something like savescumming does.

Now, if all we care about is looking at the numbers to decide who is the better player, I can see why these things seem similarly bad. But as other people here have discussed, the numbers themselves aren't really a good reflection on skill. Especially if a player isn't challenging themselves with harder combos vs. someone who is off winning with Formacids.

And as a sidenote, I think people should play a game in whatever way they most enjoy, scumming or not.

(and winning 20 consecutive games with MiBe's is still a pretty damn good accomplishment, regardless of the relative difficulty of combos)

Barkeep

Posts: 3890

Joined: Wednesday, 14th August 2013, 23:25

Location: USA

Post Saturday, 31st January 2015, 00:53

Re: Streakscumming/Winratescumming

xentronium wrote:How does this affect anyone else who plays crawl? Not in the slightest.


Agree.

Does anyone actually do this? A few people. Maybe. It seems a lot less trouble just to concentrate the effort to a single account.


Does anyone do this? A lot of people have multiple accounts, but nearly all that I have seen are accounts committed to certain gimmicks or challenge conditions or joke runs. Or sometimes just ones that people abandoned for reasons that aren't clear. (Maybe just decided they didn't like the name.)

What is the effect if many people were to do this? There would be quite a few accounts with high winrates and many wins.


It would also be hilariously, embarrassingly obvious that the winrate had been padded. Which is probably why even people who would care about having (on paper) an impressive winrate, don't actually bother doing this.

Should the practice be discouraged? No. If that's how people enjoy playing, good for them.


Agree.

What does having a long streak mean? Only that the player's combination of early game luck and overall skill were sufficient to achieve a streak of that length.


Assuming it is a legit long streak, then it shows that the person is at least very good at winning whatever combo he or she streaked with. If it is not a legit streak, it will be obvious, due to the game start/stop times, number of games played, etc. etc., all of which are logged and observable to everyone for online games.

Barkeep

Posts: 3890

Joined: Wednesday, 14th August 2013, 23:25

Location: USA

Post Saturday, 31st January 2015, 01:02

Re: Streakscumming/Winratescumming

Also I'm calling out the canard that winrates are used to justify changes to the game. I don't read the changelog and commits religiously, but I haven't seen any instances where a developer justified a "sweeping" change to the game on the basis of winrates. A few posts on the Tavern talked about overall Crawl difficulty and winrates and whatnot, but what is said here and how people on the Tavern debate the changes made to the game really is not the same as the developer's rationale. (Of course some developers discuss on the Tavern their own rationale behind changes to DCSS.)

So yes if winrate were some vital statistic being used in game design, or in something else (if prizes were given out for it, for instance) then yes obviously you would need a much more careful system. Since that isn't the case, and indeed I still haven't seen any evidence (certainly none has been presented by those most vocal about winrate-padding), I really don't see the cause for such alarm. It does read, like KL said earlier in the thread, that some people are making a big deal out of this simply as a way to diminish the accomplishments of some players, who for the most part don't seem even to care about winrates (or at least, they don't bring winrates up very often; certainly I can say that today is the day that I have most read, or typed, the word "winrate" in my time at the Tavern.)

Sar

User avatar

Ziggurat Zagger

Posts: 6418

Joined: Friday, 6th July 2012, 12:48

Post Saturday, 31st January 2015, 01:09

Re: Streakscumming/Winratescumming

Winrates in Crawl are often brought up when Crawl's fairness and randomness are being questioned. I suppose the idea that the overwhelming majority of character deaths are a result of player error and not RNG bullshit is offensive for some people.

For this message the author Sar has received thanks: 2
nago, Sandman25

Tartarus Sorceror

Posts: 1774

Joined: Tuesday, 23rd December 2014, 23:39

Post Saturday, 31st January 2015, 01:23

Re: Streakscumming/Winratescumming

I think high winrates are pretty impressive, even if it's on a new account.

Don't forget this guy who tabulated Elo viewtopic.php?f=17&t=14701 . Elo is more meaningful than winrate because it takes into account the difficulty of the characters you played, and weights recent games more heavily.
Last edited by Berder on Saturday, 31st January 2015, 01:41, edited 1 time in total.
streaks: 5 fifteen rune octopodes. 15 diverse chars. 13 random chars. 24 NaWn^gozag.
251 total wins Berder hyperborean + misc
83/108 recent wins (76%)
guides: safe tactics value of ac/ev/sh forum toxicity

Ziggurat Zagger

Posts: 8786

Joined: Sunday, 5th May 2013, 08:25

Post Saturday, 31st January 2015, 01:28

Re: Streakscumming/Winratescumming

What the fuck, Tavern.

For this message the author duvessa has received thanks: 5
Arrhythmia, Brannock, kvaak, ldf, rockygargoyle
User avatar

Tomb Titivator

Posts: 802

Joined: Sunday, 30th March 2014, 21:06

Post Saturday, 31st January 2015, 01:42

Re: Streakscumming/Winratescumming

duvessa wrote:What the fuck, Tavern.


Tavernscumming
Comborobin Admin

For this message the author xentronium has received thanks: 3
Arrhythmia, rockygargoyle, Sar

Crypt Cleanser

Posts: 723

Joined: Monday, 9th June 2014, 13:39

Post Saturday, 31st January 2015, 02:33

Re: Streakscumming/Winratescumming

xentronium wrote:
duvessa wrote:What the fuck, Tavern.


Tavernscumming

scumscumming ftw

Crypt Cleanser

Posts: 723

Joined: Monday, 9th June 2014, 13:39

Post Saturday, 31st January 2015, 18:31

Re: Streakscumming/Winratescumming

Lasty wrote:[You've asserted that manipulating winrates can be done, but you haven't given any support for an argument that it's meaningfully possible.

This obviously has upset some people so I've laid back a bit. I've tried to be really good natured about something I don't really care all that much about in the first place. I'd like to try one more time to make this clear, because it's pretty mystifying that it's not coming across that way.

I have played online exclusively since I started doing so. As I learn the game through trial and error (mostly error), these games are counted. Now, certainly, if I wanted to, I could, after winning a combo, play the next combo I am aiming for offline until I feel I have mastered it. Not only that, but I could simulate things, and even try on different armor/weapon combos with it using wizmode. Were I to do all of this before I got online to play the combo "for real," then surely you can see how that would be reflected in my winrate. If you are going to deny that my winrate would increase markedly as a feature of that behavior in and of itself, then you're simply not being rational.

So why don't I do this? It's not because I don't know how, but simply that, as a former scummer, it feels exactly like scumming. It is inorganic. It is phony and false. I don't know how else to say it. It mystifies me that this clearly scummy kind of behavior is exempt from the scummy rubric, but, just as clearly, through whatever convolution of reason, it is. Playing offline solely to bolster an online winrate is the very definition of grinding, which is defined as scumming in the design philosophy.

So how important is this in the scheme of things? Not one bit. That said, it is not isolated from the direction of game design, as stated elsewhere. While winrate may not be mentioned per se, "optimal" play certainly is, and optimal play encourages these exact behaviors mentioned.

This is no denigration of anyone's wins; that's a fallacious argument that barely rises to ad hominem. I am making a mere statement of fact about my own observations, and what someone else makes of my internal psychology from it is perfectly meaningless to the discussion.

If what I am saying has merit, then it should be addressed, either by talking about whether it is important enough to change, or by amending the design philosophy. Simply dismissing it is also an option, but a poor one. In the end, this is all I speak about, and is the reason that I speak about this at all:

Crawl Design Philosophy wrote:Another basic design principle is avoidance of grinding (also known as scumming). These are activities that have low risk, take a lot of time, and bring some reward. This is bad for a game's design because it encourages players to bore themselves. Even worse, it may be optimal to do so. We try to avoid this!


[Edit: incorrect attribution removed]
Last edited by Aule on Saturday, 31st January 2015, 19:56, edited 1 time in total.

Ziggurat Zagger

Posts: 4055

Joined: Tuesday, 10th January 2012, 19:49

Post Saturday, 31st January 2015, 18:37

Re: Streakscumming/Winratescumming

what
playing offline is grinding?
someone send help

edit: to be at least slightly helpful, in case you legitimately don't understand: you cannot throw out those offline results, just as other people have argued you cannot throw out the results on a different account. It's basically exactly the same thing as playing on a separate account (online), which everyone here has said is fine.

If you're arguing against wizmode being available, I can kind of see that argument, but then why bring up the rest of the stuff in your post?

For this message the author crate has received thanks:
Arrhythmia
User avatar

Tartarus Sorceror

Posts: 1891

Joined: Monday, 1st April 2013, 04:41

Location: Toronto, Canada

Post Saturday, 31st January 2015, 18:54

Re: Streakscumming/Winratescumming

Playing offline solely to bolster an online winrate is the very definition of grinding, which is defined as scumming in the design philosophy.


Even if people actually do this (they don't) it would be called "practice" and not "grinding", what the fuck.
take it easy

For this message the author Arrhythmia has received thanks: 2
gammafunk, rockygargoyle

Ziggurat Zagger

Posts: 6454

Joined: Tuesday, 30th October 2012, 19:06

Post Saturday, 31st January 2015, 19:02

Re: Streakscumming/Winratescumming

Aule wrote:If what I am saying has merit, then it should be addressed, either by talking about whether it is important enough to change, or by amending the design philosophy. Simply dismissing it is also an option, but a poor one. In the end, this is all I speak about, and is the reason that I speak about this at all:

Crawl Design Philosophy wrote:Another basic design principle is avoidance of grinding (also known as scumming). These are activities that have low risk, take a lot of time, and bring some reward. This is bad for a game's design because it encourages players to bore themselves. Even worse, it may be optimal to do so. We try to avoid this!

So I think a large part of this is interpretation, I believe (and it's matched my observations) that the anti-scumming philosophy applies to activites *in the game* when talking about game design I believe it's implied that you're talking about actions that take place *within the game* the game desiginers have no control over what happens outside the game, whether you take performance enhancing drugs or spend 24 hours prior to each move discussing the optimal tactic with other experienced players, or play a bunch of "practice" games offline. The rewards that they are speaking of are also *in game* rewards.

Whether you engage in meta-gaming activites that are outside the game proper, and may or may not be "scummy" isn't a subject for *game design* and *game design* philsophies don't apply to it. That isn't to say that the development crew (or the online community) approves or encourages it, but rather that it's not part of that design document, nor is it part of the things that the developers themselves are focused on.

After all, trying to influence people's out-of-game behaviors by using in-game mechanics is a losing proposition at best, and a challenge to "hackers" at worst.
Spoiler: show
This high quality signature has been hidden for your protection. To unlock it's secret, send 3 easy payments of $9.99 to me, by way of your nearest theta band or ley line. Complete your transmission by midnight tonight for a special free gift!

For this message the author Siegurt has received thanks: 3
Arrhythmia, Aule, Kramin42

Crypt Cleanser

Posts: 723

Joined: Monday, 9th June 2014, 13:39

Post Saturday, 31st January 2015, 19:06

Re: Streakscumming/Winratescumming

Siegurt wrote:So I think a large part of this is interpretation
indeed

I believe (and it's matched my observations) that the anti-scumming philosophy applies to activites *in the game* when talking about game design I believe it's implied that you're talking about actions that take place *within the game* the game desiginers have no control over what happens outside the game, whether you take performance enhancing drugs or spend 24 hours prior to each move discussing the optimal tactic with other experienced players, or play a bunch of "practice" games offline. The rewards that they are speaking of are also *in game* rewards.

Whether you engage in meta-gaming activites that are outside the game proper, and may or may not be "scummy" isn't a subject for *game design* and *game design* philsophies don't apply to it. That isn't to say that the development crew (or the online community) approves or encourages it, but rather that it's not part of that design document, nor is it part of the things that the developers themselves are focused on.

After all, trying to influence people's out-of-game behaviors by using in-game mechanics is a losing proposition at best, and a challenge to "hackers" at worst.

I agree with you completely.

Crypt Cleanser

Posts: 723

Joined: Monday, 9th June 2014, 13:39

Post Saturday, 31st January 2015, 19:08

Re: Streakscumming/Winratescumming

crate wrote:what
playing offline is grinding?

In terms of a valid, organic metric measuring actual online win rates, yes, precisely.

Crypt Cleanser

Posts: 723

Joined: Monday, 9th June 2014, 13:39

Post Saturday, 31st January 2015, 19:10

Re: Streakscumming/Winratescumming

Arrhythmia wrote:Even if people actually do this (they don't) it would be called "practice" and not "grinding", what the fuck.

At least one has already admitted he/she does, so you are incorrect that no one does this.

The rest is interpretation, as Siegurt wisely noted. You may also call it semantics.

Ziggurat Zagger

Posts: 11111

Joined: Friday, 8th February 2013, 12:00

Post Saturday, 31st January 2015, 19:15

Re: Streakscumming/Winratescumming

Aule,
Sorry, it's getting ridiculous. Unfortunately crawl is still not ideal in some aspects, it does not show very important information. Consider for a minute that AC/EV/max HP/current HP is not displayed in the game and the only way to get this information is to reproduce exactly the same character offline in wizard mode. Do you really see some honor in blind use of different body armour trying to estimate its usefulness based on your current status of "lightly wounded/heavily wounded/almost dead" vs different monsters? Do you believe those who use wizard mode would not be happy to see AC/EV/HP in the game without any wizard mode?
Now, let's back to reality. This is basically what happens with melee damage/accuracy, we have no idea how useful are some weapons unless we have very good memory and have played literally hundreds of games.
Let's complicate it more.
You are playing CSDC game, there is only 1 (one) attempt, there is no any "learning" possible ever. FoAM, tournament restriction: it cannot train skills higher than N before getting first rune. How would you play it? Try to switch to melee as soon as possible? To which weapon category? Remember - fsim before game start is useless, as it's easily possible to find 0 weapons from desired weapon category in online game before D:X. Would you stuck with your starting launcher? What should it be? Throwing/Slings/Bows/Crossbows? Remember - just one attempt, you cannot "try" them all. Should victory be determined by luck when selecting weapon category for start?
Also sometimes very funny things happen - my CeFi from another CSDC game was going to use a polearm along with longbows and got trident of distortion as its first trident (trident is really good and I had a scroll of remove curse). Good bye, dreams about using a longbow. fsim really helped me in that game, didn't it? :)

Edit. My current account has 5 games, I have never tried any of the combos (including god) offline/online. So what? Does it change anything for you? Nothing for me.

Crypt Cleanser

Posts: 723

Joined: Monday, 9th June 2014, 13:39

Post Saturday, 31st January 2015, 19:28

Re: Streakscumming/Winratescumming

Sandman25 wrote:You are playing CSDC game, ...
I was.

... there is only 1 (one) attempt, there is no any "learning" possible ever. FoAM, tournament restriction: it cannot train skills higher than N before getting first rune. How would you play it? Try to switch to melee as soon as possible? To which weapon category? Remember - fsim before game start is useless, as it's easily possible to find 0 weapons from desired weapon category in online game before D:X. Would you stuck with your starting launcher? What should it be? Throwing/Slings/Bows/Crossbows? Remember - just one attempt, you cannot "try" them all.
I remember. And that was exactly how I played it. I thought that was the point.

Does it change anything for you? Nothing for me.
Does this information change anything for you? Nothing for me. :)

Ziggurat Zagger

Posts: 11111

Joined: Friday, 8th February 2013, 12:00

Post Saturday, 31st January 2015, 19:31

Re: Streakscumming/Winratescumming

Aule wrote:I remember. And that was exactly how I played it. I thought that was the point.


I don't like to determine game outcome on turn 0 (loss with crossbows, win with slings ), not sure about you :)

Snake Sneak

Posts: 121

Joined: Thursday, 22nd May 2014, 00:57

Post Saturday, 31st January 2015, 19:40

Re: Streakscumming/Winratescumming

i tested some stuff in csdc before actually playing it

iirc tedronai did too but it didn't save him from the gnoll sergeant of doom
king of double damage

For this message the author DrKe has received thanks: 3
Arrhythmia, Aule, Tedronai

Crypt Cleanser

Posts: 723

Joined: Monday, 9th June 2014, 13:39

Post Saturday, 31st January 2015, 19:46

Re: Streakscumming/Winratescumming

All right. So here's the finish. You ready?

The failure is in the term itself: scumming. It's a patently negative connotation that can be shown to be inconsistently determined and applied, and it is unnecessarily pejorative. The reason so much ire and angst has arisen as a result of my use in relation to wins has been solely because of the pejorative labeling.

The recognition of differences in playstyles of the various members of our mammalian species without brute labeling of differences as negative can be achieved in this tiny little realm of Crawl nomenclature simply by removing one word. It wouldn't be Nirvana with hymns and hosannas, but a miniscule island of tolerance in a sea of insubstantial intolerance.

Without having once elected to be another way, it becomes all too easy to simply call another's way by curiously hateful names. I have no desire to do that, because I like to try on different ways of doing things, and I often see things in a different light when I have visited another side. Things are never as clearly defined as labels presuppose. So I propose to simply do away with the "scumming" label and all its manifestations in Crawl nomenclature. I think that would wrap everything up nicely.

Halls Hopper

Posts: 76

Joined: Wednesday, 5th March 2014, 21:07

Post Saturday, 31st January 2015, 19:58

Re: Streakscumming/Winratescumming

I think the problem is "winscumming" sounds weird as fuck.

For this message the author basil has received thanks: 4
Arrhythmia, Aule, Sandman25, Sar

Dungeon Master

Posts: 3160

Joined: Sunday, 5th August 2012, 14:52

Post Saturday, 31st January 2015, 20:06

Re: Streakscumming/Winratescumming

Aule, you seem to think that you've achieved something by redefining "practice" as "scumming". This is absurdly silly. Now we know that you subscribe to the baffling proposition that athletes runscum and foodscum for months before a big race, and most piano players musicscum for a lifetime in preparation for concerts. Heck, we've even discovered that you apparently believe that every crawl game aside from your absolute first one is scummed because you're biasing the results with your prior experience.

This is no more informative than defining "breathing" as "murder". In fact, no one could possibly believe that this is a legitimate use of language or a valid point. It's like . . .

Aule wrote:All right. So here's the finish. You ready?

The failure is in the term itself: scumming. It's a patently negative connotation that can be shown to be inconsistently determined and applied, and it is unnecessarily pejorative. The reason so much ire and angst has arisen as a result of my use in relation to wins has been solely because of the pejorative labeling.


. . . oh. Yes. So you are intentionally trolling the community and you knew this whole thing was total garbage. Great. Be warned though: trolls can get banned.

For this message the author Lasty has received thanks: 2
Arrhythmia, Sandman25
User avatar

Dungeon Master

Posts: 202

Joined: Thursday, 5th December 2013, 05:01

Post Saturday, 31st January 2015, 20:21

Re: Streakscumming/Winratescumming

How dare he ! !

Crypt Cleanser

Posts: 723

Joined: Monday, 9th June 2014, 13:39

Post Saturday, 31st January 2015, 20:24

Re: Streakscumming/Winratescumming

Lasty wrote:. . . oh. Yes. So you are intentionally trolling the community and you knew this whole thing was total garbage. Great. Be warned though: trolls can get banned.

That is so not true. This entire conversation began when I noted that winrates were not all that impressive to me, and began to elaborate on my reasons for believing so. None of what I have said in regard to why I believe that is false, misleading or untruthful for effect. You are simply being vituperative if you make such a claim.

You are welcome to call me wrong, but please do not insinuate I am a liar.
Next

Return to Crazy Yiuf's Corner

Who is online

Users browsing this forum: No registered users and 20 guests

cron
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group.
Designed by ST Software for PTF.