Spiral Knights - [STATISTICS] New Average CE per run analysis site.

[STATISTICS] New Average CE per run analysis site.

51 replies [Last post]

Tue, 11/22/2011 - 02:34

Gigitrix

======================================
EDIT: New site is online! Please use Haven-1 and update your bookmarks!

======================================
Hey there :)

I noticed there is a massive lack of information out there on hard Spiral Knights data. I intend to fix this immediately with a statistics based portion of my site!

Me and a friend started logging our raw CE gains per floor, including the floor name, depth, number of players, and Energy features purchased (Danger Rooms, etc). We pumped this all into a website where you can search for data on specific floors, depths etc, to allow others to play around with our data.

You can view the site here, but bear in mind it has no visual style yet: that's on the list. We want to expand as a full Spiral Knights fansite, but I wanted to share this information with the Spiral Knights community early, to gain thoughts and feedback.

Even with an unfinished analysis tool, on only 40 floors, we made some interesting discoveries. Some of these are already known, but are now backed up by data.

You can make a profit using CE alone, especially if you jump into the second stratum of a tier. This one is obvious, but it's still worth mentioning. Even with rising CE prices I can't see this changing soon.
For boss stratums, the "boss" isn't really the reward (tokens aside). The floors before the bosses however appear to have massive rewards. So even if you know you can't beat the Jelly King, it's probably worth your time to run that gate.
It's possible that the number of players in a group matters less than we think. NOTE: I need to do a lot more research (and get a larger dataset) before I can confirm this.

That last one is interesting. I haven't yet written the portion of the site that lets you drill down by player count, but looking at the raw data the differences are negligible. It may be worth considering that there may be advantages to running solo (you may be quicker running solo, which means more profit with the elevator pass. You will keep all of the materials found. Enemies are more predictable etc. etc.) but again I stress: it is too early to say!

Currently I have about 45 floors logged, and I hope to increase this to at least 100 by the end of the week. I will post another update on this thread with our exact methodology for measuring "profit", because I want to make sure that you are satisfied with the accuracy of our data.

Please leave me your thoughts, since I'm sure there are many forms of analysis on this raw data that we haven't even considered. We want to turn this into a pretty site with all the graphs, conclusions and ideal strategies but we need to make sure you guys are interested in it :)

View the Site

Tue, 11/22/2011 - 02:42

Trojanm

Nice work! Amt of cr also

Nice work!

Amt of cr also depends on which half of the tier is the level in. Depths after clockwork terminal pay more in cr. So for some levels it should be mentioned.

Tue, 11/22/2011 - 02:45

Ufana

Nice! "Time spent" would be

Nice!

"Time spent" would be a neat addition, too (so would be splitting no. of players into roles... but I guess that'd make it too complicated), so you can calculate "real earned cr/ce per hour" (after subtracting elevator cost).

Tue, 11/22/2011 - 02:51

Gigitrix

Yeah, it's a useless stat.

Yeah, it currently lacks a good explanation of this. The "location average credits" stat is pretty useless at the moment, but the point of that page is that you click a location link and can choose the "depth" that you wish to look at (so for instance, the "treasure room looks really sucky until you realise it was at depth 10!). In future I'll remove the stat because it's so useless, but for now I'm leaving it up as a diagnostic tool to make sure it's calculating averages correctly. Thanks for giving it a look: I've got 14 more floors worth of data to put into it from our spreadsheet so it should get more accurate over the next hour :)

It's interesting to note actually that it doesn't just seem to depend on the stratum: the second stratum of a tier obviously represents a big jump but there's a pretty significant difference even between floor 14, 15 and 16, for example, as seen on the "Depth Stats" page.

Tue, 11/22/2011 - 02:56

Gigitrix

Real Time

@Ufana We'd love to do a real "time spent" stat but it's so dependent on who's playing. I might consider having a "time=low|med|high" kind of stat based on the level name, but trying to collect "real" information on time spent would be a bit inaccurate. I do think we need to touch this in some way but I'm not just going to start a stopwatch and log times, because it's not only going to be dependent on a number of factors, but it'll make my data collection process a bit more of a rush and therefore less fun :)

I do like the idea of tagging individual levels with some kind of slowness rating though, because it is a very important factor. Perhaps I'll add user accounts and let users vote on the slowness out of 10?

Tue, 11/22/2011 - 03:06

Plinka

one question. why even

one question. why even include data for clockworks terminals? obviously you wont gain any money there

Tue, 11/22/2011 - 03:19

Ufana

I'd expect the time spent in

I'd expect the time spent in a lvl to be *mostly* correlating with "player skill" (with "equipment used" as a close second, but then again, the equipment used should strongly correlate with "player skill", too, so most of the variance would be commen), so even though the absolute average might be way off, it'd give a good idea of how a lvl stands relative to other lvls (maybe you could just standardize it to get rid of the ugly wrong time and get a neutral / artificial "lvl length" instead).

The worst interaction errors might come from things like:
- presence of vortex? (=> a lot faster if done right)
- presence of shivermist? (=> slower)
- presence of ash/tempest? (=> faster)
- How many "correct element swords" (=>faster)
- How many 2hitters (=>faster)
- How many players playing it safe / gunslinging badly. (=> slower)
... maybe subtracting the influence of vortex/shiver/2hitters_amount from the data set would already be enough to eliminate most of the unexplained variance.

(Omg sry, my statistical english is so bad and half of the terms I know seem not to have an exact translation... hope that's not written too confusing)

Tue, 11/22/2011 - 03:15

Gigitrix

RE: Clockwork Terminals

@Plinka I just put one (and one only) row in per Clockwork Terminal at 0 credits, so that this page will eventually have every depth on it. It's just a completeness/tidiness thing, and you are of course completely right. I'll do the same with Moorcraft Manor and such when I get round to it.

Tue, 11/22/2011 - 03:17

Paweu

"obviously you wont gain any

"obviously you wont gain any money there"
Actually you will, just look at the destroyable items at d23 clockwork terminal. You can also gain money from the core.

Tue, 11/22/2011 - 03:20

Gigitrix

RE: Equipment

It's certainly a complex avenue we could go down. I certainly want to do "time" as well as possible, but it's going to be a while before I add such calculations. I need to add a relative "difficulty" as well, so that people can know to avoid levels if they are less skilled/well equipped, even though it may maximize theoretical profit. No-one likes wiping out and heading back home!

I certainly think an accounts feature would be a good way to go further on. Although I'd prefer the raw floor data to come from us (to preserve it's correctness) I think there's plenty of social data about difficulty and time to derive from popular opinion.

Tue, 11/22/2011 - 03:27

#10

Gigitrix

Hmmm...

@Paweu WHAAA? I shall look into this and correct my results. Thank you for drawing this to my attention. This is exactly the kind of feedback I'm looking for, even if it's only 2 crowns, say.

Tue, 11/22/2011 - 03:29

#11

Stavrosg

Great!

Is there any way we can contribute to this?
Except during speedruns (30 minutes till reboot and we've just entered abadoned assembly? GO!(yes, we defeated RT before the reboot)) , I'd love to record those stats while delving into the clockworks.

Also, you CAN measure token rewards in the boss floors. The base profit for RJP:Battle Royale and IMF:RT is 3500cr per 15 tokens [get an armor from Binks and vendor it], or 233 cr per token, so you can add this to the level's profit.
For snarbolax it is 750cr/10 tokens or 75cr/token and for Vanaduke 30000cr/30 tokens or 1000cr/token.

Tue, 11/22/2011 - 03:37

#12

Gigitrix

@Stavrosg

>Is there any way we can contribute to this?

Not directly at this time, sorry. I don't want to open up "public" logging because I can't rely on others to log accurately, would have to fight off bots and trolls etc. We may be looking at expanding the team of loggers though some time in the future. Currently all data is sourced from me and an RL friend. There will certainly be other opportunities to help with our data collection in other areas that IS crowdsourced though, so stay tuned!

I agree that tokens have a somewhat direct relationship to reward. I feel that Tokens, Minerals and Materials are separate avenues for exploration, but it's all stuff I'd love to touch on in the future, once I've gathered plenty of crown data and have a proper site running.

Tue, 11/22/2011 - 05:25

#13

Ciopo

did you make an automated

did you make an automated tool to collect and upload the data? if so, I wouldn't mind running it to increase the sample size

Tue, 11/22/2011 - 05:33

#14

Xylka-Mkii

YAY! Finally something i can

YAY!
Finally something i can look at to know where to go!
I mean, i cant run vanaduke 24/7 without my family raging and such so with this i can judge what strata to go in with my time THANKS!

Tue, 11/22/2011 - 05:36

#15

Xylka-Mkii

Wait a minute... Is this data

Wait a minute...
Is this data from T3 and post basil?
Make sure you do some more tests, either way still very nice!
(sorry if someone stated that. im tired, gtg to school, and TL;DR the thread :/)

Tue, 11/22/2011 - 05:39

#16

Squishat

This seems interesting

I'd like to help out by logging my levels, depth and payout and posting them here for you to get more data?

Or do you need anything else xD

Either case, good work!

Tue, 11/22/2011 - 06:14

#17

Trias

Note that people have done

Note that people have done this before. You should be able to find it on the forum. People logged lots of depths. Of course, that data dates a couple of updates back (I think around the time they changed the payout for boss strata.) So it may be of limited use.

Anyway, I thought it was common knowledge that party size did not matter for the pure crown gains per level (not counting crowns made by selling mats.)

Tue, 11/22/2011 - 07:43

#18

Gigitrix

Replies

@Ciopo
did you make an automated tool to collect and upload the data? if so, I wouldn't mind running it to increase the sample size

Unfortunately no, it's all Steam Browser->Google Docs->Manual Entry at the start and end lifts.

@Xylka-Mkii
Wait a minute...
Is this data from T3 and post basil?
Make sure you do some more tests, either way still very nice!
(sorry if someone stated that. im tired, gtg to school, and TL;DR the thread :/)

The data contains all floors. It's light on T1 data at the moment, but I'm definitely going to get some from my alt. More data is on it's way every time I play.

@Squishat
I'd like to help out by logging my levels, depth and payout and posting them here for you to get more data?

There's no real public utilities we have to let people log (but this may happen in the future, with appropriate vetting and safeguards). There will be some things in the future that we need contributions on, though...

@Trias

I'm intrigued, I was never able to find any concrete data. We're definitely intending for our dataset to be long-term viable, though. If areas get nerfed/changed, we have logged the dates that the runs were made on and can archive the no longer relevant sections. The party size stuff may have been "common knowledge" but I wasn't aware, guess I should read the forums more :)

Tue, 11/22/2011 - 10:09

#19

Huhbum

fascinating...

This is incredible data, thank you for posting it.

I have one recommendation, though. I think you should re-think your approach to location data. While it is certainly true that some levels are more crown-rewarding than others(Arenas, Graveyards, etc.), the actual numbers of it vary incredibly by depth.

For example, a Depth 3 Ice Maul Arena - Robo Rampage is worth significantly less than a Depth 7 Ice Maul Arena - Robo Rampage.

Therefore, in order to make more accurate claims about locations, you would either have to compare levels at the same depth, or have data for a particular location from a series of depths to make a claim about its cr/CE comparatively with other locations. Obviously you would need more data for this, and it looks like you are well on your way to collecting more, so this issue will fix itself once your data becomes more complete.

Based on data I've collected personally(although I have less data for individual floors, as I collected information by run, not by depth), your preliminary data corresponds with mine as to your current hypotheses:

-I also have found that the number of players does not significantly affect crown gain(corresponds to the statement of others, and on the wiki)
-You can still make profit with CE alone even at the current rates, although it restricts your level choices significantly

-As to boss levels, I agree with you on the principle of RAW crowns, however in the case of RJP, the 3(maximum) tokens you gain, if converted into raw crowns via the 15 tokens -> Brute Jelly Mail series(3500cr sell value) automatically adds ~700cr worth to that depth, which you may or may not want to take into consideration.(edit: someone already mentioned this earlier! my bad.)

Again, this is a great site and I look forward to your continuing results.

Cheers.

Tue, 11/22/2011 - 10:44

#20

Culture

Hmm

I agree with the other posters. If you can defeat the boss then definitely do it. The tokens have value when sold to a vendor. Many people have no idea and go up when they get to the boss. Your dataset doesn't have to reflect this however since it isn't a direct CR reward.

I'm glad you are doing this project. There are some questions I hope the data can answer:

Should I wait 15 minutes for the arena to cycle back around? What is the spread at this depth for clockworks vs arena?
How much better is RJP or FSC vs a non-boss dive?
Is any particular strata theme more profitable?
Considering just crowns, is it worthwhile to open a Danger Room at my current depth?

I think that eventually the data will need to be grouped by depth or strata. That way you can compare a depth 24 arena vs clockworks, for example.

Tue, 11/22/2011 - 11:43

#21

Gigitrix

Replies and order switching

@Huhbum
>While it is certainly true that some levels are more crown-rewarding than others(Arenas, Graveyards, etc.), the actual numbers of it vary incredibly by depth.

This is certainly true. Although the data quality is fine (especially if you click ALL on the location) I think the way I've set it up so that you pick Location before Depth is wrong. I'm going to reverse this so you see, say, "Depth 17" before picking an appropriate location that we have logged.

Token stuff certainly needs to be acknowledged. Once we have more data on the total number of locations etc we want to write descriptions of them, which will be a quick and easy way of acknowledging factors like tokens, difficulty etc. into consideration before I start approaching these areas with a more statistical frame of mind. Additionally I might make a toggle switch so that you can remove boss floors from averages (since they are obviously going to be lower than their "true" value).

@Culture

I intend to reverse the filtering to "Depth->Location" from "Location->Depth", as well as work toward using the data to interpret answers to these questions. Of noteworthy mention is that we are collecting information on danger rooms/loot rooms/energy based features, it's just that the public interface doesn't yet allow users to drill this down. The data collection continues, as does our site development :)

Tue, 11/22/2011 - 12:43

#22

Chris

I've been doing things like

I've been doing things like this for my guild Knightmare for almost 13 months now :)
This is going to piss off OOO as people will be making more profits and will be wasting less ce ;)
Sent you a friend request in-game in regards to possible collaboration.

I think I speak for all of us when I say we greatly appreciate your work.

Tue, 11/22/2011 - 17:33

#23

Kwizzy

I have a ton of extra entries

I have a ton of extra entries I can share with you if you'd like.

I have a similar site (private, very ugly!) that I & friends have been using, but we didn't think the forums folks would be interested. My past profits per level posts never got many responses on here. And it kinda removes some of the fun of Spiral Knights having it all public IMO.

Once you get a sizable # of results, you can easily predict how much a stage is worth on any floor. Clockwork tunnel style & Graveyard levels have a much larger variance than the others, and treasure vaults use a different equation (chests scale a lot less than monster kills)

You might want to include tokens too! The random drop ones are worth 100 cr each (or more). Number of people in a group really matters for profitability in stratums 4-6 - the materials are usually worth more crowns than the levels themselves.

Is any particular strata theme more profitable?

Yup! I'll leave that as a mystery ;)

Considering just crowns, is it worthwhile to open a Danger Room at my current depth?

Assuming you won't die & enjoy the fights, it's worth it at any depth -except- the very first one in that tier. Especially if someone else is paying the energy costs hehe. A bit of Gun Puppy, my favorite one to farm, for example:

Depth - cr - Tokens (not counted in cr)
1 - 114 ***
1 - 152 **
1 - 115 *
2 - 264
2 - 234 ***
3 - 302 ***
3 - 240 **
5 - 358 **
5 - 484 **

Many of the other types don't even get 100 cr (not counting tokens) on the first stages.

Tue, 11/22/2011 - 17:33

#24

Jamesburger

Good Job

Nice site. Good Job!

Tue, 11/22/2011 - 20:57

#25

Algol-Sixty

First, thanks for publishing

First, thanks for publishing this data, it is useful. Many people have collected similar data, but rarely is it organized and published as well as yours.

One comment though. You seem to repeat claims in your post and on your website such as: "You can make a profit using CE alone, especially if you jump into the second stratum of a tier. This one is obvious, but it's still worth mentioning. Even with rising CE prices I can't see this changing soon." Your data really doesn't back that up. For example, the average T2 run pays 742.89cr per level, which is barely break even at the current cr<->CE exchange rate. And that average is only that high because you include a lot of what are thought to be the most profitable floors, such as arenas and the JK and it assumes that you never have to energy revive.

I don't think anyone has claimed that you can't earn a profit running the clockworks if you are good enough to almost never need an energy revive, you have enough friends to be able to skip most of the odd level strata, and are in T2/T3. However, that really cuts out a large percentage of the player base, especially new players.

Your data gives some good insight into the situation, but by making such strong claims that aren't backed up by your own data, I worry about the bias in your data.

I think it would be more productive to explore what it takes to break even, rather than make bold claims.

Thu, 11/24/2011 - 05:16

#26

Gigitrix

Sorry for the delays (replies)

Apologies for the delay in getting back to people: I've had a busy schedule the last couple of days. Finally have a full day I can devote to the site which is great :)

@Chris
Excellent! I'm always interested to see how people react, and if we can remove effort duplication then that's excellent. Even if not, I'm sure we can learn from your experiences :) I feel only the "hardcore" types are going to be that swayed by the information to try and maximise earnings, so I doubt OOO will be fuzzed too much! And if they are, they'll just have to add some new areas for us ;)

@Kwizzy
That sounds excellent, you've put a lot of effort into posting, thanks. To answer a few things:

I'm surprised you thought people wouldn't be interested! It's certainly an aspect of the game that's fascinating, and although it does reveal information, I highly doubt it's spoiling that much of the gameplay. Just because the boss tiers are most profitable doesn't mean I will always do them, for example :)

We certainly want to include variance etc. as other statistics that are public. It's easy done, but hardly useful right now! I've noticed similar results with regards to treasure rooms but it's such early days for my dataset at least.

Tokens have been discussed and I certainly need to address them eventually. The thinking is that the current data represents raw credits, but I want to collect further info on mat. drops and tokens, possibly with some semi-updated database on market values for end products. Then I'll be able to display "Raw Crowns","Goods value Crowns" and "Total Crowns".

@Jamesburger
Thanks, hope you like it :) I am planning on fleshing it out today, so that it can sort through data in different a
ways, and isn't quite so ugly :)

@Algol-Sixty
It's completely not my intention to make misleading claims. I do think it's possible to profit from jumping in to the second stratum, as I stated earlier. My Depth stats page illustrates this, and that's without counting any alternative income streams like tokens/mats. HOWEVER I completely agree that it is not my place to make such bold assertions: the data alone should do the talking. The claims I made were an attempt to justify looking at an ugly looking "table of numbers" when I thought no-one would be interested due to it's poor layout. I accept as well that it represents the "late game" somewhat. I'll remove this off of the site to avoid any confusion, and because I don't want to appear like I'm trying to sell some "uber-1337 secrets of Spiral Knights" or whatever :D

Obviously this data reflects perfect runs with no revives. We certainly want to factor in subjective information like Time taken and Difficulty in the future, but it's outside of our very limited scope at this early moment in the site's history :)

Thanks for your very legitimate concerns and well thought out post. I certainly want to add more sections of the site with more information about our methodology and so on, so that our methods can be critiqued and improved if needed.

Thu, 11/24/2011 - 06:15

#27

Trias

Sortable tables.

I think it would be a huge help, if it were possible to sort the table on your site. One way to do this is through javascript. See for example: http://www.kryogenix.org/code/browser/sorttable/ . Put there are other ways to do this depending on the infrastructure you use.

Thu, 11/24/2011 - 06:16

#28

Gigitrix

Sorting

@Trias
It's on the list, thanks for the plugin link :)

Thu, 11/24/2011 - 09:38

#29

Hanktron

just a few thoughts

I probably won't be messing with these data, but I took a look at them and had a suggestion or two.

If I do end up playing with them, I'd really like a raw spreadsheet download. This would allow me to filter and sort in Excel (actually, I might end up using R, but that's beside the point) to my heart's content.

I'd also love to see the addition of a few columns. I see you have a cr/ce column, but I think it'd be nice for a simple "CE Spent" column. It wouldn't be hard to retroactively add this for runs that have already happened. The data should exist somewhere since you have to make the calculations anyway.

Also a stratum flavor column would be nice as well. With enough effort, one could figure it out but I think a simple stratum type column would make some (very interesting) analyses a snap. It'd be annoying to retroactively do this, but for all future data entries it wouldn't be tough.

Thu, 11/24/2011 - 10:43

#30

Gigitrix

@Hanktron

CSV import is live here, I'm just testing it now.

CE Spent is coming once we start adding frontend stuff about Paid Energy features used: it's in the db, it's just not ready yet for public consumption as there's a lot of work to do with regards to figuring out which features were used. I might do a "3CE*features used" cheat in the meantime though, that's not a bad idea and wouldn't take much work.

Stratum flavours presumably affect which levels are generated (as seen here), so I'll be storing them more as properties on each location than on every log entry: it's rather awkward to figure this information out per floor. You'll then be able to hopefully filter by "Stratum: Fire" and it will just treat it as a group of locations. I'll be able to add stratum comparison stuff too: the raw data needed is already there: it's just the metadata on locations that needs adding.

@Everyone
I changed the way the site frontend works: you can now filter by any number of combinations of Players, Locations and Depths. Create your own reports and have fun with it, but remember comparing raw "Player Count" is misleading: you are only looking at the logs we have, without comparing "like for like", so if we did lots of solo tier 1 runs it's going to look like Tier 1 is massively unprofitable, when it could just be an issue with the logging. Please compare using depths for now if you want meaningful results :D

Sun, 11/27/2011 - 05:57

#31

Gigitrix

News and sorting!

@Trias sorting is now in. It doesn't look like it, but you can click the column headings and sort ascending and decending. I REALLY need to get the site designed properly.

Oh, and we've surpassed 100 logs now. (YAY :D) I really need to make the navbar counter dynamic. I'm still far from happy though. We need to make the results statistically significant, or at least approaching it...

In other news, planned features:

Compare results. Compare, say "1 Player" with "4 Player" on a given Depth, for instance. Adds more deduction power and flexibility.
Energy Gate based features. Compare "No Gates" with "Treasure Box opened" with "Danger Room Opened".
Community account system and submission process. This WON'T be us letting everyone log results, but there's a ton of data around the raw logs that people can help with (and can be demonstrably verified and removed if erroneous, unlike floor logs). We MAY be looking into letting some trusted peers have access to a logging system though, but we will have to speak with you in game/over steam first and there needs to be enough interest to make this worth the effort. Should this feature go ahead we'll add an optional override to view only our data if you're not convinced by our vetting and training process ;)
Subjective (voted) data collection, like "difficulty" and "speed"
Programmer friendly .xml and .json feeds for pages, IF there's demand and people can think of a need.
More Spreadsheet Friendly .csv renderings of pages, for those of us who still prefer our analysis via spreadsheet
Pretty graphs!
A Useable Site Design!

Sun, 11/27/2011 - 11:36

#32

Mohandar

Suggestions

Some thoughts:
1) Instead of using ce gained, use cr/e spent. So if you gained 1300cr but spent 10e + 3e for an energy gate, then your total for that level is 100cr/e. Because CE changes in price all the time, this is probably a more useful measure.
2) Classification by level type in addition to depth. So, arena, clockwork tunnel, decon zone, aurora valley, jigsaw valley, etc. If you wanted to get fancy you could include status type, but that seems to be less important.
Edit: Oops, my bad. Didn't notice you'd already added level classifications, great!

The way I can see this working out is to have a front-end where trusted users can submit 3 (possibly 4) pieces of information: crowns gained, depth, and level type. The optional 4th is tokens gained (and what type), since all tokens have a hard lower bound conversion to crowns (100cr for sparks, 233cr for jelly, bark, 75cr for fangs, and 1000cr for seals).

I would be happy to contribute, as would most of my guild (Delicious Bacon). Send me a mail in-game with details, or catch me online sometime.

Mon, 11/28/2011 - 17:00

#33

Gigitrix

@Mohander

@Mohander
Instead of using ce gained, use cr/e spent.

It's literally in development right now. We have carefully logged instances where CE is spent, it's just not on the "frontend", which is something I need to correct. Our data distinguishes between "Danger Room" gates, "Loot" gates (where there is just boxes) and "Other" gates (usually extra areas of the levels) so you should be able to do some comparisons on a more general level with these to see if it's "worth" opening these gates. Development is slower than I'd like due to coursework but it's definitely coming!

I know you've seen our level based stuff but I want to take it even further: if we could input the raw level data on the gates every two days we could figure out roughly where the best rewards lie. It's probably going to be unsurprising though, it'll most likely be the boss tiers. But it might help those who are sick of grinding the bosses but still want to maximise crowns, so they can pick #2 in terms of profitability.

Token stuff is a separate feature and it's something I'd like to explore along with mats, but I really want to concentrate on getting solid info on the crowns first. It's pretty easy of anyone to view existing data and interpret it for things like tokens anyway, and if anyone's stepping up to the challenge I'd be happy to provide any of the outputs from my tools in a programmer friendly output so that token/mat stuff can be added in by third parties.

Mon, 11/28/2011 - 20:00

#34

Anakletos

helpfulness to newer players

I find this rather interesting, and will probably follow it for a while (maybe contribute if I can remember).

That being said, can you make the rest floors a little more noticeable? I tend to think "I am two floors beyond Moorcraft" more than "I am at depth 10." It would be easier for quick referencing/alt+tab. For example, places like Moorcraft Manor and the Clockwork Terminals could be colored or in bold. As of now, I can only find them because they are the ones with numbers near 0 in the first two columns.

In addition, IDK about other people, but I never go from the depth 1 to the depth 29 when running through the Clockworks. Highlighting terminals and special places would make it easier to find your starting position.

Mon, 11/28/2011 - 20:44

#35

Gigitrix

@Anakletos

@Anakletos
Great Idea: I added background colours when you filter by depth to correspond with stratums and tiers. Green=Tier1, Yellow=Tier2, Red=Tier3, with the darker variant corresponding to the second stratum of that tier. Thanks for your suggestion!

Tue, 11/29/2011 - 02:02

#36

Gigitrix

Volunteer Signup/Initial Registration

Many apologies for another double post, but you may be interested. I'm going to collect in game names and emails of those who may be interested in helping us to log results for the project. It's early days yet, but when we are ready we can then contact those users and quickly grow the number of floors logged on the site. I'm hoping to build a nice utility that will allow for easy external submissions after a manual approval step, which should eliminate any chance of foul play (log entries carry a user ID anyway so it can be traced back if questions are raised)

The long and short of it is, if you think you might want to help the site by logging your crown yield per floor, please drop your details in the form below and we'll send you an email when we're ready. You're not committing to anything, we're just trying to see if it's worth our time building a user submission system at this early stage. I may contact a couple of you in game who've already expressed an interest earlier on in this thread, if I see you online.

https://docs.google.com/spreadsheet/viewform?formkey=dEpBTEJ2TFU3ZjJ6cUF...

Tue, 11/29/2011 - 13:36

#37

Antistone

Problem

Took a quick look, I notice you've logged a depth 17 graveyard with 94 crowns. That's...quite low.

I suggest that you only log levels where the level was completely explored and all loot collected (except perhaps stuff behind energy gates, and maybe with special entries for arenas with only 1 or 2 fights completed). I wouldn't necessarily worry about destroying every single torch, but based on my experience, completely clearing a stratum 4 graveyard should easily yield in excess of 1000 crowns, and if you're logging runs where players skipped 90% of the level's loot, your data is going to become meaningless pretty quickly.

You should also clarify whether your totals include crowns from the prize wheel spins. Keep in mind that you can receive crowns from party members' prize wheel spins, not just the one you see, so if you want to exclude them you need to write down your crown total before you step on the elevator.

Tue, 11/29/2011 - 14:24

#38

Xylka-Mkii

Nice!Finally came back and

Nice!
Finally came back and the data looks good!
Mind if I add in some statistics I got?

Here it is:

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Depth17
Ice Maul arena- Cadaverous Clash-----2269 cr-----times done 1 ||||| EDIT: We did the full 3 fights in the arena and cleared all torches, boxes etc.
----------------------------------------------------------------------------------
CR intake in FSC on average
Blackstone Bridge-------1016 cr---- Tested 1 times

Charred Court------------
8 rooms ----- 1958 cr----1825cr second time-----Tested 2 times |
|6 rooms done ---1384cr---times done 1
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

I know its not organized too well, but hopefullly you can get it sorted! (^^;

Thanks again for the stats!

~Xeirla

Tue, 11/29/2011 - 15:54

#39

Gigitrix

RE: Problem

@Antistone

That's very odd. I've looked at the raw database and I think I found the problem: "Graveyard - Where Monsters Fear to Tread" is location ID "17" so I think I must have transcribed this record wrong. I've deleted it from the database. This is clearly erroneous: about 3 enemies would yield that much at that tier. Many apologies: I'd just like to add that this is unlikely to happen again because I now have a tool that will import data from my spreadsheet automatically. Thanks for spotting it.

As to our methodology: prize wheels are excluded from the results and we take great pains to ensure that the "start" crowns is recorded at the start of each floor (not just assumed from the previous floor) and the "end crowns" just before the next lift for the very same reason. Fortunately both of us are on fast computers and are usually the first to load, so in cases where we're in groups we can record the proper "starting crowns" amount before people start to smash the couple of vases/torches in the starting area.

We explicitly only log fully completed areas (yes, including smashing every vase/uncovering every grass. I consider that the stats are most valuable to Free-to-Play players, for which time isn't a major factor to optimize for (energy is), so it's important to me that we get this right. We've had to throw out a few firestorm citadel floors because people weren't going down the alternate routes after saving the required number of sprites to move on.

In short, we take our data pretty seriously, but we apologise that our full methodology has been pretty opaque up until now. We want to write a full document explaining absolutely every aspect of this (to allow people to scutinize and improve our methods, as well as allow more hardcore users to get on board and follow our protocol) but we never got around to it.

So unfortunately @Xylka-Mkii we're unable to use your data, because we have to make totally certain it's recorded in a specific, almost algorithmic way :/ It means we have far less floors worth to work with, but the accuracy we get in return is, in my opinion, worth it!

Tue, 11/29/2011 - 16:12

#40

Bloody-Nermy

good work

i just want to say good work :D

Tue, 11/29/2011 - 19:21

#41

Xylka-Mkii

@Gigitrix aww QQ. I even made

@Gigitrix
aww QQ. I even made my cr count absolutely 0 for each floor too...

ah well, either way keep up the hard work!

Tue, 11/29/2011 - 19:31

#42

Xylka-Mkii

I filled the questionare,

I filled the questionare, hoping to help out officially!

Sat, 12/03/2011 - 14:06

#43

Asukalan

amazing site

Sun, 12/18/2011 - 18:15

#44

Gigitrix

Nearly done now!

Hey all, sorry for the lack of progress! I don't know if anyone saw from my notice at the bottom of the page, but I had to put the project into "3rd year of uni coursework shutdown" mode because of my poor time management skills! I have 2 more days worth of work to complete and then I can finally boot up Spiral Knights again and start working on the project (including my partially completed AWESOME PAGE DESIGN! More logs will be going up soon (just in time for Christmas, in fact!) and I'll hopefully be coordinating external contributions soon as well!

It's going to be a statistics filled holiday season, and you're all invited!

Sun, 12/18/2011 - 19:23

#45

Kalaina-Elderfall

I was logging crowns a while

I was logging crowns a while ago before I got my elevator pass. If you're interested in my data, I uploaded it here:

https://docs.google.com/spreadsheet/ccc?key=0An2wgNSdMWQZdEN2Z01hV2ZBN2p...

Column I has the crowns earned. In cases where I opened a danger room, it's Clockworks+Danger as the type (never opened treasure gates). Profits don't include materials or minerals. Dunno if I was recording everything you were recording, but I figure I'd post it in case it's useful.

Sun, 12/18/2011 - 21:24

#46

Realnight

Looks like a few things got

Looks like a few things got misplaced?

http://www.tobypinder.com/site/spiralknights/crownData/13/all/filter

Sun, 12/25/2011 - 20:44

#47

Quailia

wanted to bring this back to

wanted to bring this back to the front page, cause I want to see this sort of thing expanded on in the community.

Also, does anyone happen to know what their gain on the tier 1 section of the new JK gate that opened as of today? It has that post-basil arena at d7 that is very appealing, and I'm wondering if I can use it to continually play, buy ce, play, etc. due to the new lower CE prices.
Edit: because I'm only on mist/ce, and am feeling cheap enough to not buy the new pass. not to mention that having the pass I suspect would absolutely ruin my life :X Also, I'm specifically looking at grinding tier 1 because it's easy, and I can potentially pick up the heat from nooby noobs :)

Mon, 01/09/2012 - 12:43

#48

Orzy

<3 it :$

Sat, 01/28/2012 - 23:50

#49

Gigitrix

We're not dead!

Remember when I said I was working on a site? Well my uni exams are over, and HERE IT IS!

Haven-1

Currently it lacks a few features like the ability to sort column headings and such (everything from the old site will be ported over, don't worry!). It's a complete rewrite of the code so that it's more durable, and a pretty nice Spiral-themed design.

Additional good news is that I'm currently ironing out kinks in a "process" (not code) that will allow me to take submissions from trusted invididuals, whilst verifying these results. This should increase the amount of data we've collected DRAMATICALLY, with the side effect of letting us have the ability to explore item/material drop rates as well.

As always, let me know what you think, but I'll be making a new thread soon to properly announce it once I've added more content (we won't just be doing data collection: there's other stuff planned for the site)

Sun, 01/29/2012 - 00:51

#50

Traevelliath

Overall

I like the new layout a lot, its a lot more aesthetically pleasing IMO.

As for usefulness... I would really like it if the sort feature thing would have an option like "Firestorm Citadel" in addition to each Firestorm Citadel, instead of each individual level. Then maybe a total of all the FSC levels for overall gain/loss. Most people think of the second half of each tier via their respective boss stratas. Tier 1 - Snarby. Tier 2 - JK/RT. Tier 3 - Vana. Most people don't think of them as the individual floors of each boss strata.

Search

[STATISTICS] New Average CE per run analysis site.