Final Phonebook Import
Sep 19, 2016
Over the weekend, I did a final import of the phonebook and voicemails.
Here are the stats:
Total names in phonebook: 3457
Total voicemail messages: 1848
Never entered a name: 457 (not included in counts above)
Total size of all messages (uncompressed): 2.72G
Total size of all messages (processed, compressed): 166M
Of the 1848 VM files, 7 were uploaded corrupt and couldn't be converted (or even listen to). Since it was a small number, we'll email those people to get us new files.
3 people have incomplete information. It looks like they started, but never finished. They will also get an email.
27 backers went to the webpage, but never entered information. Over the course of the past 6 months, we've sent out a lot of reminders, so they will not be included.
We have not done a vetting of all the VM to see if any are obscene or violate the rules. Of the few hundred I've listen to so far, they all are fine. I'm optimistic that very very few (if any) will be culled.
There is a drastic range of recording quality (as you'd expect on a real VM), so some of them might be hard to understand. They were all normalized to try and get a consistent volume. I'm sure we could spend a lot more time processing the recordings, but we also have a game to make.
Thanks to all the Kickstarter backers who submitting voicemail messages. There is an achievement for listening to all 1848 of them. Have fun with that.
- Ron
I'm speechless!!!
On same topic... what's your opinion on achievements for games?
Just imagine you finally get green tentacle the record contract, and then there it says: "Achievement unlocked: green rocks!".
I'd already be distracted from enjoying my own success.
Btw, Thimbleweedpark meets ClassicComputing in Germany yesterday:
http://martinwendt.de/twp_meets_caren.png
Greetings from PriorArt =)
Ah, memories...
The Kickstarter ended nearly two years ago and a lot can happen in such time span. I personally lost two close relatives in the last two years and I know of at least one backer who won't be able to enjoy the game:
Astrid Beulink aka Pollo Diablo passed away just shortly after the Kickstarter: http://www.adventuregamers.com/articles/view/27948 :-(
Challenge accepted.
What a fantastic opportunity you've given us all to feel even MORE like a community!
You guys are WONDERFUL!
http://ik2.dk/phone.wav
Unfortunately it turned out that is was violating several rules and as I really didn't want to listen to my own voice in the game I never made a new one.
/Kim
I'm curios about the voice messages the backers came up with, I will definitely call some.
Another voice related question: Do you have any plans to synchronize the lip movement with the voices of the characters?
I stumbled upon the rhubarb lip-sync project, which looks great! There's a demo video on youtube:
https://www.youtube.com/watch?v=OX_K387EKoI
I really think it would add a lot to the experience.
He wrote that it would be a matter of how much time they have.
I love to hear some step noise, pulse repeater, 2600 bursts, crosstalk, vintage busy, reorder, ring back, SIT etc.. And if you call to Germany in the 80's, there was no system 5 over here. You most likely would hear some Siemens motor switches instead (like those on the Florida recordings on that archive).
I think you mean "normalized," which makes the volumes consistent... "equalized" is where you change the levels of different frequencies (add/subtract bass, mids, highs, etc.)
It's like if someone has food stuck in their teeth, better to let them know, it's not a big deal, just point it out.
If you wanna go on we can too :)
The game resources should be at the frequency which is native to the game engine to avoid unnecessary resampling. Of cause, the samples should be processed with a bandpass filter. Only frequencies between 300 Hz and 3500 Hz should be there. Those were filtered quite steep in order to modulate multiple carriers per lead and demodulate them at the distant office, similar to AM broadcasting. You could occasionally hear high frequency hiss on long distance calls, since the filters weren't perfect.
Due to the narrow band usage, those samples would compress quite a lot better then hi-fi samples nevertheless, hence it can probably be encoded the same way, any other sample in the game are encoded as well.
That's my whole shtick, you're putting me out of business here!
But, it would be nice if the beeps (if any) are not louder than the speech. Unfortunately, everyone has his/her own sense of hearing.
Either by providing a final code that people send in via anonymous post card or some direct uplink to the TWP-Masterbrain-Mainframe that obviously exists?
Or something inbetween? I love feedback on minutes of playtime or number of room-changes or items as long as it is somewhat voluntary.
Think about this: If you have an analysis on how long a player needed to solve a certain puzzle you still don't know if the player had fun trying out the different things he/she had in mind or was frustrated while trying to crack it.
It might be useful on crashes - but other than that: What's the point? Is there a Thimbleweed Park 2.0 with all the glorious number-based improvements from the data collection? Is any of that "knowledge" really that transferrable to a nother game that it justifies the time spent for analyzing the data?
One of my favourite examples on why blind belief in numbers is a bad thing: Microsoft uses meta data collection on a grand scale to "optimize" their operating system and especially the user interface. Their numbers told them to get rid of the start menu in Windows 8 (besides a lot of other dubious improvements). It's something the tooted around and were VERY proud of. And we all know how well that turned out.
Bottom line: Either you are a good game designer or you are not. Numbers aren't funny. Numbers aren't emotional. Numbers aren't memorable.
The information is useful in two ways: 1) For future games, just understanding play patterns is good. 2) If we find a unfair and hard spot in the game, we can fix it and patch.
Still that is super-helpful. We see how many rooms people passed (i.e. how much they run back-and-forth) and also average playing time (which even for such a low number statistic is a very nice 'log-normal' distribution :) as well as flags for individual optional puzzles. We don't judge total numbers such as playing time as success or failure (after all, tax computations take a long time as well) but it is indeed a great tool to estimate how good our estimates are : )
Can you reveal what project you're referring to? I'm not in the games industry, but this concept of measuring some statistics, and asking people to send back small counters that help provide feedback seems quite good for most software development.
http://martinwendt.de/caren/
A very cool adventure for your C64
Can you reveal what project you're referring to? I'm not in the games industry, but this concept of measuring some statistics, and asking people to send back small counters that help provide feedback seems quite good for most software development.
Actually I understand your point of how worthy such meta data can be /might be. But I see this data worth mainly from a research point of view. So if a product is well tested you actually must not need this data. Right?
Nevertheless, as we can see at the Steam Global Gameplay Stats for MI2, most people who finished act 1 completed the whole game. I assume that, if all of them were stuck due to a particular puzzle, most of them would look in the internet for a walkthrough after a while, so that they would be able to continue (and finish) the game. In order to avoid players to blame the game for a too difficult puzzle, you should simply check out whether all puzzles are logical enough, so that players would have to blame themselves.
By the way, modifications are not necessarily advantageous (see the classic Star Wars Trilogy). At least, I would say that any changes should be as inconspicuous as possible.
That's all I'm talking about. There are some "optional" paths through the game that don't trigger achievements, and I'd like to know how many players take those. It's very high level. I would like to know how much time players spent with each character.
I'd like to know the time between play sessions. Did players complete the game in a few closely spaced sessions, or did players go away for weeks and come back. How long did it take players to complete the game? Players personal recollections are often suspect. This is all very useful information. Maybe not for this game, but certainly for my next game.
Why wouldn't you want this information? I don't know a game designer working today that doesn't want and collect this type of information. It's fascinating and we talk about it all the time. Data collected on someone else's game is useful to me if it calls out play patterns, even if it's not an adventure game.
Maybe you could also ask for additional and optional personal information (e.g. number of kids explains those many weeks of gaming absence between sessions :-).
I'm not saying we will do analytics, I honestly don't know if we have the to implement it. We'll probably just have to rely on achievements to track usage. Do achievements bother you? You can't opt-out of achievements and they are tracking you as well and worse, they aren't anonymous like analytics are.
I buy most of my games on GOG (or similar places) because I despise DRM (and I don't see it as my job [anymore] to fix games which are broken by definition).
And since I don't use GOG Galaxy I actually don't have achievements most of the time unless a game also implements them "offline", e.g. like Shovel Knight does (they call it Feats).
I use Steam very seldom although I do have quite a lot of games there too (Steam keys are often included with DRM-free purchases like Kickstarter or Humble Bundle).
My last real purchases on Steam which I can remember were The Cave and Costume Quest 2 (for both I had no hope for proper DRM-free versions, but I wanted to play them anyway).
Just don't implement analytics in a way they are required or blocking the game/UI when there is no network.
They destroy the (next) game. Let me explain that: Games are art, they are made in a creative process. The best games I have ever played were made by small teams, which developed that game as they like - and not the players oder companies. They haven't used any analytics.
If you begin to develop a game based on analytics, you make the game "easier". You cut allegedly difficult parts. You "soften" the whole game. At the end you make a mass product. You make another "grand war of battlefield auto". Why shouldn't have a game a difficult puzzle? Why the game shouldn't have some puzzles that only some players will see?
Yes, Monkey Island 2 had that monkey wrench puzzle. But all the world is still talking about it. If it hadn't the monkey wrench puzzle, it wouldn't be Monkey Island 2. Yes, Zak had mazes. But now it's known for the mazes. (Back that time we discussed how to master the mazes! That was fun! Yes, really!)
Data analytics are like smoking: If you begin with it, you can't stop anymore. You want more and more. And at least the results destroy the (next) game. Don't do that. Be creative. Don't smoke data.
And one last thing: We are not the kids with unlimited time anymore. We don't search about alternative paths through a game. And if we get stuck in an adventure we are using a solution. So your analytics are lying.
From the developer's point of view, MM and Zak were not only adventures but also ventures, because they were in uncharted waters back then. Albeit, they have become quite popular.
Well, I wouldn't demonize analytics, but, when you collect informations, you have to take care in drawing the correct conclusions. If the statistics tell you how popular a playable character is (as Ron mentioned), it's much easier to reasonably decide on the cast of playable characters for a sequel. In DOTT, for instance, some players might have missed their favorite character from MM.
Being stuck over a too long period can be very frustrating, because you wouldn't like to have to look for a walkthrough in order to make progress. You rather would like to cope with it yourself, so that you can give yourself an honest pat on the back when you've finished the game. Nevertheless, when you play an adventure game (or a puzzle game in general), you usually expect it to be not too easy. Personally, if I only need one or two afternoons in order to finish an adventure game I never played before, I would probably find it to be a bit too easy and I would be disappointed.
By the way, the monkey wrench puzzle has definitely not been the reason why I like Monkey Island 2. In my opinion, it wouldn't have needed any analytics in order to recognize that the monkey wrench puzzle was too far-fetched. On the other hand, I like MI2, despite the monkey wrench puzzle! So, statistics can't tell you everything. Your intuition and your gut feeling are not less important than analytics.
"Kids, don't smoke data!"
Everyone is already using Analytics. There is nothing new here, it's be going on for years.
I would never use analytics for this. This is what bad marketing departments do.
Anyway, I'd love to know ... could you actually basically watch someone play the game, with those analytics? They sound like *everything* is being watched, and in my head this is starting to look like you're basically recording it all, like you could in MI with shift+10 if I remember correctly.
Really not a fan of these analytics, for a reason that is difficult to explain, and may be a little nonsensical. I enjoyed playing games like Zak, Maniac Mansion and so forth, -- it was like a little, living, breathing world completely *disconnected* from the outside world. I found that very, very fascinating on a technical level. It's still basically just like that -- a self contained world - with rules and characters and a story, and all.
You see? I told you I wouldn't make much sense trying to explain it. When this is constantly connected in some way, shape or form, I also just feel watched somehow. This is basically why most (all) programs don't even get to go online here. Except the activation of course.
Did I mention I'd love a DRM free GOG version of any game? If not for backup purposes, then for "non-reasons" like this?
I find this discussion of Analytics fascinating. It's so common in games and game designers love this stuff and live by it, but it feels like it's something that is new to the blog readers.
There seems to be this belief that they hurt creativity. Nothing could be farther from the truth. No one good makes decisions based on analytics numbers. I would love to know what the popular Maniac Mansion characters are. I can guess, but that data only comes from vocal people on the internet. I'd love to really know. What if I found out Wendy was the second most popular character when playing Mansion Mansion. That would surprise me, but it's useful to know, not that I'd run out and give her her own game or something silly like that, but it's just interesting and might make me create a future character a little different. Creativity needs a lot of inputs and knowing how people play your game is one of them.
"No one good makes decisions based on analytics numbers. I would love to know what the popular Maniac Mansion characters are. I can guess, but that data only comes from vocal people on the internet. I'd love to really know. What if I found out Wendy was the second most popular character when playing Mansion Mansion"
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Please, can you tell me if this is intentional? I mean, not always, just now. I have the impression that this time it was genuine :-)
Yep, he does it sometimes, see https://twitter.com/AdrianJacek/status/735527490526011392
And thus Edmund Mansion mansion was born!
Hmm... Maybe I didn't make myself clear.
Yes, I remember that tweet. That was the first, genuine one. I find it lovely that he repeats sometimes that typo on purpose.
It was funny, people liked it, and it became an inside joke. But I had the impression that THIS time, it was genuine. You know, the same mistake again. We'll never know :-)
You invented a diabolical puzzle, but you discover looking at your analytics, that most player solve it in 5 minutes. Is it because they are clever, beacause your puzzle is far less diabolical than you thought, or because they look up the net?
What you mentioned (80% of players stop playing at the same spot) is not a problem that should be solved after the game was released but in play testing. Also, in reality, nobody would get stuck today on a puzzle for too long before he/she looks the solution up. Some people might Alt+Tab out of the game for this. A lot would probably just use their smartphone - which is invisible for your data collection.
But even if you think you can get useful information out of the collected data. You still have to interpret the results. And that is HIGHLY subjective. In other words: You see what you already think you see.
Going back and fixing part of the game play because of analytics is something I have never seen anywhere ever. These methods are only used in highly dubious games with micro transactions to streamline the game, place pain points and optimize willingness to pay. In other words: These tools are used by assholes. The only reason game designers of REAL games go back to previous work to change it is because of massive fan outcry. And when they do there is normally an almost as big outcry from the people who do not want the game to be changed.
You say yourself you wish you had analytics for the development of Monky Island. But why? They are the most highly regarded adventure games of all times. What could you have POSSIBLY improved with analytics in those games? The monkey wrench puzzle? Let's assume you HAD the analytics tools at that time and the players HAD internet connection: By the time you analyzed the data and came to the conclusion that a lot of non-english players had problems with that certain puzzle the knowledge on how to solve it would already have spread amoung player (remmber: in this scenario there is "an" internet - and therefore also forums and help sites). So you would be fixing something that is somewhat broken but that the players already have adapted too.
The only reasonable argument for data collection in games is if you are planning a series of games (like the afore mentioned FIFA games) or collection technical data for crash analysis.
You have posted yourself in some of the others replies that you wouldn't use it for creative decisions - good! But what for then? Technical assistance - Okay. But what else is there? Play patterns? And what would you do with analytics about play patterns? As you have admitted yourself - you can't possibly know from the data if the player had fun or not. Changing the "patterns" because of statistical data is just as good as guessing because you cannot predict if more or less people will like the "improved play patterns" in the next game.
Even on the business side of software where I earn my money we cannot trust data analysis. One example: We ran an analysis over all customer installations to see which of the advanced features customers used how frequently and used this data to hide less used features (we made them accessibly over a "Add feature" button). BUT: The data cannot show us if the customer didn't use the feature because he/she didn't need it or because he/she didn't find it in the first place or didn't know about the feature. So maybe we are now hiding some things that our customers want but couldn't find well enough in the first place. And now we are just making things wore because we are burrying it even deeper.
I only know of negative examples when it comes to data-driven design.
Besides the example with the start menu, here is a different one: Microsoft once decided to hide the less used menu items in it's office products (I think it started around Office 97) and would only show them once you hovered your mouse on a "down" arrow on the bottom of the menu. A LOT of people that we had contact with in customer support at that time were completely irritated because the function they wanted to use wasn't there anymore. Yes, the didn't use it very often. But that made it only worse! Because they didn't use it often they were already unsure of how to find it. And then Microsoft burried it even deeper.
By the way, if Monkey Island had analytics, maybe Ron would have recognized how many people had bought (or pirated) his game back then. Maybe he wouldn't have left Lucasfilm, in this case, until he had done a third episode. :)
Very true.
But I think many designers often miss an important part here: never change an interface if you haven't to (and don't change more than absolutely required). Changes are always costs and people avoid costs!
The GUI is as well an interface, even more it's a very special one cuz it's IT-human and connects two different "systems". So one has to recognize two things: don't change without need and respect both end of this interface for its own!
Costs for humans are time to adjust. And nobody likes this if there is no need (advantage perceptible).
Microsoft changed the GUI much more than needed (they always do for new releases) - first fail.
And imho they didn't respect at least the human side - also fail.
a) known for relying heavily on telemetry and data analysis when it comes to designing their software
b) the producer of windows (and after him many other people involved at MS) at that time explicitly said this was the reason for removing the start menu.
Read it up if you want: http://www.alphr.com/news/enterprise/375550/why-microsoft-killed-the-windows-start-button
In our case the code also acted as proof of having completed that particular version. It also detects pirated versions :)
https://ksr-ugc.imgix.net/assets/005/263/658/acb8217e55d20c856e0db59f910eb81e_original.jpg?w=680&fit=max&v=1453962257&auto=format&q=92&s=ffde83f896f8c2aecc25dab55c86ddfd
:-)
(I hope the link keeps working, otherwise click my name and search for Caren)
Unfortunately my extreme OCD means that I feel a need to try to 100% games I play
Achievement should be enjoyable, not something that would tempt me to mute the game for an hour + while i watch a movie on another monitor and systemically click on name after name.
Too many achievements like this has caused me to skip playing games entirely in the past.
Guys
Guys
Call me 🤙
7483 xxx