Forum:Charm log limits

From the RuneScape Wiki, the wiki for all things RuneScape
Jump to: navigation, search
Forums: Yew Grove > Charm log limits
Archive
This page or section is an archive.
Please do not edit the contents of this page.
This thread was archived on 5 April 2012 by Thebrains222.

Currently on charm logs we require there to be at least 50 kills for a submission, but why? What's wrong with someone giving say 20 or 30 submissions? It's still accurate, assuming it isn't vandalism. This is one example of a good submission that was removed simply because it has under 50 kills. Monsters, like mith dragons or GWD bosses, that can only be killed a small amount of times per trip, for most players, aren't going to get many submissions above 50 kills, so shouldn't we also accept less kills, when they are just as useful?

I'd suggest making the lower limit somewhere around 10-15, as that is enough to be fairly accurate, without being an excessive number of kills.

Discussion

Support - As nominator. Hunter cape (t).png Sentra246Blue hallowe'en mask.png 09:47, February 4, 2012 (UTC)

Comment - In the example mentioned, it can be seen that the user found no blue charms on his run, despite there being a small percentage of chance of receiving one. What if, under the proposed changes, many users simply do 15 kill logs but continue to receive 0-1 blue charms? Then the percentage of blue charms would be driven down, and the average percentages may become inaccurate. Chicken7 >talk 10:04, February 4, 2012 (UTC)

What if users also constantly get 2 blue charms? It would then rise. Overall it would balance itself out, as having 3 people submitting 20 kills each, would be the same as 1 person submitting 60 kills. Some people will get more, some will get less, but again, overall it will balance itself out. Hunter cape (t).png Sentra246Blue hallowe'en mask.png 10:14, February 4, 2012 (UTC)

Support - A user can kill 500 rock lobsters and get 0 charms. This is all about chance, so limiting the number of kills for a submission is silly if the reason is inaccuracy. sssSp7p.pngIjLCqFF.png 16:32, February 4, 2012 (UTC)

Slight support - I agree with what you're saying but I think the proposed figure is too small. I think 20 might be a more ideal number, as it's less likely to be affected by awkward numbers. Under the current proposal, if somebody submits a charm log of 10 and gets a blue charm, that's 10% of all charms blue, even if, as is the case with the mithril dragon, the average is 4%. Now if it was 20, that single blue charm would account for 5% of the charms, a closer figure to the average What I've done Ciphrius Kane Talk 19:54, February 4, 2012 (UTC)

Support for higher level monsters - Mithril dragons, sure. Goblins, definitely not.

The reason we have a kill limit is because of human nature. Users are a lot more likely to remember a trip (and hence upload it here) if they got something significant, be it a rare drop for a drop log, a blue charm from a cow, or whatever. If we allowed log entries for trips of any number, we'd be introducing biases. A person who kills 10 cows and gets a blue charm is a LOT more likely to log that than a person who kills 10 cows and gets nothing, despite the second situation being much more likely.

By imposing the charm entry limit, we both increase the chance of getting something memorable, and we weed out the people who killed 5 monsters for the heck of it, got something good, and want to log it. --LiquidTalk 20:02, February 4, 2012 (UTC)

Even for low level monsters, 50 is still overkill. If someone put in 20 kills that would still be pretty accurate. Plus what's stopping someone who got 1 blue charm from 50 kills and then deciding to log it because it was good? It's exactly the same as someone with 20 kills doing the same. Hunter cape (t).png Sentra246Blue hallowe'en mask.png 01:11, February 5, 2012 (UTC)
As I've said before, with more kills you're more likely to get a significant drop. --LiquidTalk 03:14, February 5, 2012 (UTC)

Support/Alternate - Instead of a hard limit, how about something to the effect of 50 divided by the number of charms per drop (since that seems to be linked somewhat with the number that someone might kill in a trip/couple of trips), in the case of Mithril dragons, which drop 4 at a time, 50/4 = ~13 (12.5 but rounded up), Kuradal apparently assigns 5-35, so the majority of Kuradel Mithril dragon tasks would be acceptable. (Duradel only assigns 4-11).

Essentially, change the policy to: rounded up/down value of 50 div by number of charms at a time, discretion by RC watchers/undoers of a further +/-10% or so (why invalidate 45 kills, when it's pretty much 50 anyway).

Can any statisticians provide any evidence/reasons that this might not be such a great idea/wouldn't work? --RSDaftVader 04:12, February 5, 2012 (UTC)

Support for high level monsters - I would see this as making monster drops for high level monsters more accurate, as a large number of smaller submissions can be better than a small number of large submissions. Requiring less kills opens the charm drops to players who killed a smaller number of the monster. However, I don't see the need to reduce the number needed for lower level monsters. If we reduce it too much, we seriously risk the accuracy of our reports (after all, I can go in, kill a single monster, get a blue charm drop, and despite that being pure luck, it's a 100% chance, technically). Hofmic Talk 07:19, February 5, 2012 (UTC)

That wouldn't reduce the accuracy at all. Every kill is independent, so if 100 different people submitted only one kill it would be an accurate sample of 100 kills, so long as someone didn't choose certain kills and ignore others. The risk of accuracy is for different reasons to that. --Henneyj 23:35, February 10, 2012 (UTC)
Except, as someone already pointed out, people have a tendency to remember notable drops more than unnotable drops. They're often more likely to submit drop rates if they obtained note-worthy drops than if they had a dry streak. Though, you're right, besides that. I just worry that the number of submissions wouldn't be much higher, which could offset the reduced kills per submission. Hofmic Talk 06:44, February 17, 2012 (UTC)
Was only only clarifying for that one specific point not in general. --Henneyj 02:41, February 20, 2012 (UTC)

Oppose - As far as I know, a bigger sample space means more accurate data. With this in mind, the larger kill requirement of 50 makes more sense. Although a reduced kill requirement may reasonable for higher level monsters. I believe that overall, it is in the best interest of the Wiki to have the 50 kill requirement for greater accuracy of data. I am open minded...convince me otherwise. Raglough 08:14, February 5, 2012 (UTC)

Support - I think some people are missing the point of why we have the limit in the first place. It's not for the sake of having a higher sample size; in fact, having a higher limit actually decreases the total sample, because people who would have otherwise contributed to the log would not do so because they didn't have the 50 kills needed. I also don't think we should have different limits for different monsters, as it's confusing to the people who add to the charm logs. I think a minimum of 20 for all monsters would be good. These are not rare items like godswords or Torva pieces; they're charms, and even the rarest charms aren't that rare. The problem of people updating after they get a couple good charms is overblown. Besides, if we seriously cared about the accuracy of the charm logs, we would split them by level. ʞooɔ 09:07, February 5, 2012 (UTC)

Partial support - Drop logs for 50 is indeed to many for larger creatures harder to kill, however, these should only apply to those like really hard monsters, i.e. Nex, or Zilyana, but for something like metal dragons, or gargoyles, perhaps change the setting to the lowest slayer amount for that type of monster, as more than likely, it is those that are on their slayer assignment to keep track of how many they killed and have them at once. -- Dragon longsword.png Cire04 TalkAttack.png 02:20, February 11, 2012 (UTC)

Support for high level monsters - This would be good for higher level monsters like bosses, but 10-15 kills is a little low in my opinion for regular monsters. Full slayer helmet.png Axela40 Slayer-icon.png 18:11, February 12, 2012 (UTC)

Oppose - I'm quite swayed by Liquidhelium's point. The actual quantity of kills in the submission is not necessarily the real issue; instead, the problem comes from a bias toward recording memorable drops rather than everything one kills. I'd agree that people are more likely to remember to log a drop when they receive one or more "rare" charms, and what that could lead to is a blue-bias in the statistic. Let's assume that we remove the kill limit, on the principle that any limit is arbitrary and counterproductive. If we buy into the assumption that people are more likely to record rare drops, what we'll likely see is a slow drift toward a higher proportion of blue-crimson charms on npc's than we had originally. For every person who records an entire slayer assignment of cows (say 50) and receives a normally proportioned set of charm drops (1 gold, 2 green, 1 crimson, 0 blue), there may be several individuals who will record only when they receive a rare charm, or a predominance of rare charms because they feel it's worth recording. If 5 people record these rare drops (at 10 kills each), you've now got a total of 100 kills with say, 6 gold charms, 12 greens, 8 crimsons, and 5 blues. That works out to a very different drop rate than what usually would be recorded. Even worse, we can't just revert drop rates which deviate from our pre-existing drop rate because the deviation may be due to Jagex changing the drop rate, or because our original rate was actually wrong... we can't necessarily be certain that our existing rate is truly correct. I would fully support a decision to remove the limit if we could be reasonably sure that there wasn't some sort of rare-bias out there, or that we could accurately identify such a bias. Penderwyll 19:38, February 12, 2012 (UTC)

Oppose - But I do support using common sense and allowing <50 drops for high level monsters that aren't killed in high amounts, or monsters that are not killed very often (charm logs with <1000 kills are still quite common). For monsters like Goblins, submissions of <50 kills will likely just become too inaccurate to be kept, and there are enough others submitting >50 kills to make it no problem not to keep those submissions. For the higher levelled (I was thinking of cb 200+) or less frequently killed monsters, I think a minimum of 10 or 15 would be better. If this would become the final outcome of this thread, I am able to change the Drops submissions script according to it, so that is not an issue. JOEYTJE50TALKpull my finger 12:58, February 16, 2012 (UTC)

Semi-support - for lower-levelled monsters, 50+ will prevent people from rushing to add an entry every time they get a blue charm. For higher-level monsters though, say anything assigned by Duradel, Lapalok or Kuradal, the minimum submission should be the lower number out of either 50 kills or the lowest number of that monster assigned by those slayer masters or, for mithril dragons, the lowest assigned by Kuradal (as Shilo assigns only a handful at a time). Small recharge gem.png AnselaJonla Slayer-icon.png 13:08, February 16, 2012 (UTC)

I won't be able to get the amount of monsters assigned by the slayer masters in the drops submissions script, because there is no information about that on monster pages. I can only do anything with information on the Charm: page for the monster and data that is already on the infobox on the monster page itself. That means I can check if the monster is assigned by certain slayer masters, but I can't check the amount of assigned monsters.
By the way, Duradel assigns a minimum of 4 Mithril dragons, and it would probably be a bad choice to allow that. JOEYTJE50TALKpull my finger 13:30, February 16, 2012 (UTC)

Support for high level and infrequently killed monsters - I have often wondered why we have the limit. Even if 50 people submit 1 kill apiece, that's still the exact same as 1 person submitting 50 kills. This won't change much for frequently killed monsters with logged kills in the tens of thousands, but could increse samples drastically for very high leveled or uncommonly killed monsters. For example, I'm sure more than 758 [[Bat/Charm log|Bats]] have been killed; logs such as these would become much more accurate. Guthix raptor.png Soy the Stig or to friends, Xolecon Dark cavalier.png 21:54, February 16, 2012 (UTC)

Fyi, the amount of bat kills submitted is currently 4,089. JOEYTJE50TALKpull my finger 08:43, February 17, 2012 (UTC)
Charm log reading fail. My bad. Point still stands though. Guthix raptor.png Soy the Stig or to friends, Xolecon Dark cavalier.png 21:24, February 17, 2012 (UTC)

Oppose - The "why" is that using lower limits results in inaccurate numbers. People tend to focus on whatever they consider abnormal or different (in general, not just with charms). As such they are more likely report their charms if their current task has a different proportion of drops from the current reported amount. They are less likely to report if their charms are about the same as the current reported amount. By requiring larger samples, the error in each report is smaller, and the error over all samples is also smaller. This is very standard sampling practice in any field that uses statistics and needs to discourage anecdotal data from skewing the results. By decreasing the required sample size, we'll mostly accumulate a large collection of anecdotal data, which may or may not average to the actual drop rates. --Saftzie (talk) 23:22, February 16, 2012 (UTC)

Sure more kills will make it more accurate, but 50 is really overkill. Surely it could be even more accurate having say a minimum of 20, and getting more submissions? While I say that, allowing smaller submissions isn't going to dramatically increase the number of submissions we get, in fact I doubt it will increase by much at all. Hunter cape (t).png Sentra246Blue hallowe'en mask.png 09:03, February 17, 2012 (UTC)

UCS - For a monster like Nex, no one will really record 50 kills, so that can be lower. But for something like a guard, 50 seems better. If it's a high level monster, I don't see why the limit cannot be lowered a bit. ɳex undique 01:18, February 19, 2012 (UTC)

I agree, we don't really need to be really strict about the number of kills, and we should make exceptions. Ork legions are a good example - you can only really kill four a day. The fact alone that we have a guideline helps to reduce lower submissions, even if the guideline isn't always enforced. --Henneyj 02:39, February 20, 2012 (UTC)

Bump/request for closure - No discussion for a month. I don't think anyone has anything else to say, else they'd have done it in the past month. JOEYTJE50TALKpull my finger 08:46, March 23, 2012 (UTC)

Closed - Charm log submission limits will remain at 50 kills for most monsters. Monster who are difficult to kill in large numbers will have their submission limit lowered to 15 kills. We can determine whether a monster is difficult or not with common sense. 222 talk 10:10, April 5, 2012 (UTC)