Difference between revisions of "Talk:Judge"

Revision as of 16:34, 17 September 2009

Deletions

Props to you guys for working out a proper rating system!

I have issue, though, with the deletion of the "Excellent weapon" pages: They're useful in that many players will only want to see the weapons in the 15th category of quality. Even if we scrap the pages in question, there needs to be a simple way to separate the "best" weapons from the not-so-great ones.

If we proceed with the deletion of the "Excellent" category pages, I propose that we add a second sortable table to each page, including only the 15th quality category.

-TherionAndAlts 06:51, 16 September 2009 (UTC)

Why would we need a second table? Sortability means that there is a simple way to separate out the best weapons: sort by rating and only look at that chunk of the table. ETA: And there's a good reason not to have two tables: unnecessary redundancy. It'll be two things to update on that page every time you add an excellent weapon (or change its judge info, or give it a more precise weight). I'm actually ambivalent about having those tables there at all, for that same reason, since they have to be updated by hand. You can't just add a weapon article; you also have to go to its category page and update that. It's very useful to see them all in one place, yes, but I think the right thing would be to have the tables generated automatically... I'm just not sure if that's even possible here. Anyway... there's also the question of where to put the cutoff. If a weapon has a rating of exactly 94 or 95, is it excellent, or not? Anyone with a high bonus will get "extremely good". Would that mean that if you don't have a high enough bonus, you can't put something in that category? --Ilde 09:45, 16 September 2009 (UTC)

The second table allows people to sort only "excellent" weapons by weight, specials, etc. You can use the main table, if you like (that's what it's there for), but the second cuts out the dross such as butterknives and rubber chickens, which many players don't care about because they're not "excellent." I often use the second tables for that reason; I care about the weight of a tiger fang relative to a kodachi, but not the kodachi relative to a copper short sword. (Insanity actually requested that we expand the second table, to include "Extremely good" weapons as well!)

If you, or anyone else, can work out a way to automagically generate tables filled with article data, then I'm all for it! The current tables are the way they are because I wished to add sortability but am inept with Wiki code. The current tables are not hassle-free, but I feel that the small amount of extra work involved is worth the return in convenience until such time as an automagical table can be generated.

As to the exact cutoff, I think we can work that out when we find someone with a high enough ot.ev.we bonus, and enough patience, to determine exact ratings for each and every weapon. In the meantime, I'm happy to have a broad range; since the "excellent" tables are now "100 club" tables, we can use the broad range of, say, 94-100, since that is the range in which judge will have an "excellent" result.

-TherionAndAlts 12:34, 16 September 2009 (UTC)

Hrm. As I see it, there are essentially four main problems with having a second table, by whatever name:

• Some weapons will need to be added twice - once to the 'normal table', once to the 'special table'. Ilde has already gone over this point above, so I won't elaborate much - essentially the issue here is that it's then easy for the information in both tables to get out of sync.
• Where is the cut off? - It looks like everyone wants this at a different level!
  • It's currently (roughly) a rating of 94+
  • Insanity says he wants 'extremely good' too? Well, that's anything from 81+ to 89+.
  • Perhaps we want round numbers - 80+ or 90+?
  • Someone really picky might only want 98+
• You can't determine rating accurately enough (or at least, you can't using judge). Say Alice, with ad.ev.we bonus 400, judges the 'Womblefork' weapon (which actually has a rating of 95) as 'Extremely good'. She would enter the weapon has having rating 89-95. Now Bob, with ad.ev.we bonus 163, judges the 'Froghammer' weapon (which actually has a rating of 94) as 'Excellent'. He would enter the weapon as having rating 94-100. Most 'elite weapon' tables will now end up including the Froghammer, but not the Womblefork - despite the fact that the Womblefork actually has a higher rating.
• Is rating the thing you really want to use to distinguish weapons?. Rating is a fairly generic measure, which is used in the judge command to give a single representation of the weapon's quality. It's based purely on the weapon's average and max base damage, and isn't always the most appropriate measure. Consider now the following weapons:

Name	Min dmg	Ave dmg	Max dmg	Rating (derived)	Weight	Difficulty
Foobane	50	100	150	60	5 lb	40
Frogcleaver	110	110	110	57	4 lb	50
Blade of Fluff	50	105	160	64	10 lb	75
Gunfork	0	90	180	62	3 lb	20

Alice the Assassin is off to kill giant spiders. She's probably best off using Frogcleaver here, as average damage will be most relevant to her attacks. Were she to choose a weapon based on rating, however, she'd go with the Blade of Fluff, which would actually be less effective.

So should we order on average damage instead? No. Bob the Barbarian is off to kill Troll Grflxen. A Troll Grflx has very tough skin that absorbs 160 hitpoints off all damage. Choosing either Blade of Fluff (best by rating) or Frogcleaver (best by average damage) would be foolish - poor Bob won't damage those Troll Grfxen at all! He should go with the fearsome Gunfork here (best by max damage).

Polly the Priest is now off on a jaunt through Ankh-Morpork, killing a pretty random selection of things (with varying levels of armouredness). She wants a good all-round weapon, so rating is probably more relevant to her than either max or average damage alone. She therefore chooses the sturdy Blade of Fluff. But wait! Would she be better off using the Foobane instead? Sure, it does slightly less average and max damage, but it weighs half as much and is easier to use. She won't hit as hard using it, but she'll hit more often, and be less likely to miss, and that will probably more than compensate.

The moral of the story here is, we should probably just present all the data that is available, and let the viewer decide how to order it.

Given the above issues, I think we're probably best off just with having a big sortable list in the categories - if the viewer wants to see the 'best weapons', let them sort the table according to whatever criteria they desire.

On a related note, I think what we actually want for weapons in general is to install either the SMW or Data extensions - that way auto-updating tables can be created.

--Chat 16:04, 16 September 2009 (UTC)

Ooooo! I like the look of SMW better--this page looks like exactly the kind of thing we want, whereas I don't see an example of anything like that for Data.

--Ilde 20:17, 16 September 2009 (UTC)

The current tables are not hassle-free, but I feel that the small amount of extra work involved is worth the return in convenience until such time as an automagical table can be generated.

Oh, I definitely agree there. It's miles better than not having the tables and it seems like the best we can do with the wiki as it currently is. It just offends my delicate sensibilities to, basically, have it as a table instead of a view. ;)

--Ilde 20:17, 16 September 2009 (UTC)

Keep in mind that the second table is only linked from the main weapon page; the players that don't want to use it don't have to. It's there for the ones that do.

The cutoff is not really as relevant to me as the ability to sort by multiple conditions. Since I don't know how to make a single table that allows you to sort primarily by rating region (I propose 95-100, since iirc that's the widest range that can possibly show up as excellent) and then sort that by weight, the second table (which shows the weapons that fall between the example of 95-100), is useful.

If you know how to make a table sortable by multiple parameters, then I'd be all for the single table; but as it is, we don't have such a method. Once we do, the second table can be dropped with no complaints from me.

Regarding your example, is there actually a way to accurately determine the base damage of a weapon (without knowing its exact rating), its difficulty, and the relative power of its specials? If an accurate and objective method exists, I'd like to add that to the infoboxes and tables; they would be a great way to differentiate the weapons.

I use the rating to sort weapons because, as far as I am aware, it is the only objective way to determine a weapon's quality in terms of hard data.

It's also worth pointing out that if we install the SMW or Data extensions, the problem of out-of-sync tables are solved, and the only remaining issue is the cutoff.

-TherionAndAlts 00:08, 17 September 2009 (UTC)

This seems to work: sort by the least important attribute first, then by the most important attribute. So, sorting by weight and then by rating will mean that, within a group where the rating is the same for all of them, records are sorted by weight. --Ilde 01:41, 17 September 2009 (UTC)

How exactly did you achieve this? If it does work, then we've solved out problem. -TherionAndAlts 01:47, 17 September 2009 (UTC)

Exactly the way I described. First click the thing at the top to sort by weight, then click the one to sort by rating. I assume that when it sorts two records that are the same with regards to the column you're using to sort, it just leaves them in whatever order they were in. --Ilde 01:55, 17 September 2009 (UTC)

Excellent, though unless I'm mistaken, that only solves the problem so long as the ratings are displayed as broad ranges. Once we cut the ratings down to tighter ranges and eventually accurate values, the problem will arise again (because the ratings being so precise will render the multiple-sorting function essentially meaningless).

Why not change the "Excellent swords/maces/etc" category to "Swords/maces/etc with a rating of 95 or greater?"

-TherionAndAlts 02:11, 17 September 2009 (UTC)

Regarding your example, is there actually a way to accurately determine the base damage of a weapon (without knowing its exact rating), its difficulty, and the relative power of its specials? If an accurate and objective method exists, I'd like to add that to the infoboxes and tables; they would be a great way to differentiate the weapons. I use the rating to sort weapons because, as far as I am aware, it is the only objective way to determine a weapon's quality in terms of hard data.

Well, the main article describes exactly such a mechanism, though you will need to get a rating range first. The only other way to determine these characteristics would be to work backwards through the steps at the weapons article - have one PK hit another a huge number of times and record each damage, get a min/max/average from that, then factor out your skills, the effects of critical hits, enchantment, weapon condition, armour, etc. Given the huge number of variables involved, that would be practically impossible to do with any accuracy (especially when you try to factor out the critical/exceptional/marginal hit effect), so I suggest sticking with the mechanism from Judge.

As for difficulty, I'm afraid that's something that definitely exists in the code, but can't very feasibly be measured by players - for that you'd need to get two PKs to work backwards through the to-hit process, which is way more complex than even the damage calculation.

Determining the power of specials is something that's only going to be possible with research - the mudlib has all the guild command code (including all specials) removed from it.

Excellent, though unless I'm mistaken, that only solves the problem so long as the ratings are displayed as broad ranges. Once we cut the ratings down to tighter ranges and eventually accurate values, the problem will arise again (because the ratings being so precise will render the multiple-sorting function essentially meaningless).

Why not change the "Excellent swords/maces/etc" category to "Swords/maces/etc with a rating of 95 or greater?"

I'm afraid that barring the exhaustively long process above, there really is no way to get 'accurate' or 'precise' ratings above 95 - the best you can possibly get from judge are the rating ranges as presented at 392 bonus - raising your bonus above that doesn't actually improve the accuracy any further. The issues with categories called '<weapon>s with a rating of 95+' are as per my post above - you can't accurately determine the rating, and the cutoff is arbitrary.

--Chat 21:19, 17 September 2009 (UTC)

Specials

It seems you don't always get told which specials are available with a weapon... perhaps this depends on whether you have those specials? When I judged Sarilak with a character without any specials (but an ad.ev.we bonus of 192), it didn't tell me the specials even when I succeeded... but now I'm judging it again, with a character who only has 188 ev.we (but who has a lot more specials available) and I'm seeing that it can be used to pierce and slice. --Ilde 01:55, 17 September 2009 (UTC)

I propose having that character learn commands, one at a time. Judge Sarilak after every new command, and if the specials appear only after you learn them, then there's your answer. :) -TherionAndAlts 02:12, 17 September 2009 (UTC)

Ha, I guess that would do it... now I just need to find a character who doesn't have specials, and that I don't mind putting a bit of xp into sp.we and ad.ev.we. --Ilde 02:33, 17 September 2009 (UTC)

We can also do a generic test- Smugglers get ad.ev.we as a primary, but not being Warriors, do not have access to Impale, Behead, or Crush. If a Smuggler judges a weapon known to have any of those specials, we can get our answer quite easily. I'll ask one of the Thieves to do it when next I'm on. -TherionAndAlts 02:56, 17 September 2009 (UTC)

Did that. Feef has a 324 bonus and judged a rose-hilted long sword; he did not see that it had impale or behead, whereas I, as a warrior, know that it does support these commands. -TherionAndAlts 03:06, 17 September 2009 (UTC)

Yeah, the mudlib says that you'll only see a special when judging a weapon if it's a special that you already know. --Chat 21:22, 17 September 2009 (UTC)

Wait a moment...

I just reread the Judge article, and came to the conclusion that judge is not subjective, provided that the player is skilled enough. The article says that at an ad.ev.we bonus of 392, a player can achieve a completely accurate result.

If this is true, and I am interpreting the table correctly, then any weapon with a confirmed rating (taken with a bonus of 392+) of 96 or greater is without question an "Excellent" weapon.

Is this correct? -TherionAndAlts 12:53, 17 September 2009 (UTC)

Careless wording on my part, I'm afraid. The way it actually works is that 392 is as accurate as you can get - increasing your bonus above that won't improve your accuracy. At 392, your accuracy still won't be precise. You're still going to have ranges for ratings.

If it's accuracy you're after, the best way to do that would be to get two people with different bonuses to co-operate on judging. For example, if player A (bonus 150) judges a weapon as 'Rather good' and player B (bonus 250) also judges it as 'Rather good', then that weapon's rating is in the 56-68 range, and the 68-76 range. Therefore, its rating is exactly 68. Sadly, this trick won't work for weapons of rating over 95 - there's no range overlap there.

I suppose you could argue that any weapon judged as 'Excellent' by someone with 392+ bonus is one that everyone (or at least, everyone capable of judging that weapon) would necessarily also judge as 'Excellent', but I still see that as a rather arbitrary distinction for creating a separate category.

Anyway, this is mostly a moot discussion, because currently no one has a rating over 392 - the highest is Greyko, and he's only got 331.

--Chat 21:34, 17 September 2009 (UTC)