Failure rates of Roguelike Games
Jeff Lait
development. I shall first present some dubious statistics and then
you shall complain that they don't accurately reflect roguelike
development. Complainers may find some solace that my own roguelike,
POWDER, has just fallen off the 6 month cut off to be an actively
developing roguelike. The shame! The shame!
To find the previous four studies, search for Failure inside this
newsgroup.
The data for this comes from:
http://thelist.roguelikedevelopment.org/
which Michal Bielinski has been maintaining.
Infinite thanks to Michal for maintaining that list, or we would not
have another year of meaningless statistics to bore you with!
This years data comes from June 25th, so is 5 days earlier than the
usual July 1st data snapshot time.
3 51! 52!
2 ^ ^
1 # # #
0 # # #
9 # # #
8 # # #
7 # # # #
6 # # # #
5 # # # # #
4 # # # # # #
3 # # # # # #
2 #### # # # #
1 #### ## # # #
0 #### ## # # #
9 #### # ## # # # #
8 #### #### # # # # #
7 #### ###### # ## # ##
6 #### ######## ## #### # ##
5 #### ######## ### # #### ## ###
4 #### ######## ### # # #### # # ## ###
3 #### ############## # ###### # ## ###
2 ##################### # ###### #########
1 ##################### ######### ##########
1 000000000111111111122222222223333333333444>+123
1 123456789012345678901234567890123456789012>
This tracks the number of roguelikes by last release date. The first
column has a # for every roguelike released in the last month. I have
omitted the last column which would have all the roguelikes over 42
months old or without known release dates. There are now 123 such
roguelikes being tracked.
The peaks at 4, 16, 28-29, and 40 month marks are due to the 7DRL
challenges. The peak on the 35th month is due to the 1KBRL
challenge. There is an interesting spike at the 10th month that
corresponds with last years ARRP.
Note that the last two year's 7DRL challenges have caused spikes of
over 50 roguelikes! To conserve USENET bandwidth, I have cut off
those bars.
Next, we will look at the cumulative totals for the last year.
Copying from the last seven year's reports:
Numbers: (July, 2011)
Month # Total Percent
1 21 21 4%
2 12 33 7%
3 12 45 10%
4 51 96 20%
5 2 98 21%
6 6 104 22%
7 9 113 24%
8 8 121 26%
9 11 132 28%
10 14 146 31%
11 7 153 33%
12 8 161 34%
Rest 308 469 100%
Numbers: (July, 2010)
Month # Total Percent
1 14 14 4%
2 7 21 6%
3 14 35 10%
4 57 92 27%
5 10 102 29%
6 5 107 31%
7 3 110 32%
8 1 111 32%
9 6 117 34%
10 1 118 34%
11 1 119 35%
12 1 120 35%
Rest 223 343 100%
Numbers: (July, 2009)
Month # Total Percent
1 6 6 2%
2 16 22 8%
3 7 29 11%
4 17 46 18%
5 8 54 21%
6 6 60 23%
7 5 65 25%
8 3 68 26%
9 5 73 28%
10 3 76 29%
11 14 90 34%
12 6 96 37%
Rest 166 262 100%
Numbers: (July, 2008)
Month # Total Percent
1 11 11 6%
2 5 16 8%
3 6 22 11%
4 12 34 17%
5 14 48 24%
6 6 54 27%
7 1 55 28%
8 5 60 30%
9 2 62 31%
10 2 64 32%
11 4 68 35%
12 2 70 36%
Rest 127 197 100%
Numbers: (July, 2007)
Month # Total Percent
1 10 10 6%
2 6 16 10%
3 9 25 15%
4 11 36 22%
5 9 45 28%
6 5 50 31%
7 5 55 34%
8 3 58 36%
9 3 61 37%
10 2 63 39%
11 1 64 39%
12 2 66 40%
Rest 97 163 100%
Numbers: (July, 2006)
Month # Total Percent
1 9 9 7%
2 3 12 9%
3 3 15 12%
4 11 26 20%
5 5 31 24%
6 1 32 25%
7 2 34 26%
8 3 37 29%
9 1 38 29%
10 3 41 32%
11 4 45 35%
12 2 47 36%
Rest 81 128 100%
Numbers: (July, 2005)
Month # Total Percent
1 15 15 15%
2 3 18 17%
3 10 28 27%
4 12 40 39%
5 2 42 42%
6 1 43 42%
7 5 48 47%
8 2 50 49%
9 3 53 51%
10 2 55 53%
11 3 58 56%
12 2 60 58%
Rest 43 103 100%
Numbers: (July, 2004)
Month # Total Percent
1 6 6 10%
2 5 11 19%
3 2 13 22%
4 3 16 27%
5 0 16 27%
6 0 16 27%
7 4 20 34%
8 0 20 34%
9 0 20 34%
10 1 21 36%
11 2 23 39%
12 2 25 42%
Rest 24 59 100%
The increasingly meaningless Percent Actively Developing Roguelike has
maintained its position despite what would seem an inevitable trend to
zero. This year the 34% of tracked roguelikes were updated in the
last year. This is largely due to the enormous success of this year's
7DRL challenge, but note there is an equally large number of
roguelikes weighing down the percent from last year's challenge!
More interesting is the absolute number of touched roguelikes. 2006
seems to have been an anomaly as we've continued to see growth in this
area with a record 161 roguelikes updated in the last year.
This chart shows the number roguelikes touched in the last 6 months,
12 months, and the percentage the twelfth month number comprises of
the total number of roguelikes being tracked.
The New column records the increase in total tracked roguelikes. The
assumption is that these roguelikes were added in the last 12 months
since the last report, but it is possible that older roguelikes were
added to the list that inflates this number.
The last column is Old. These are roguelikes that have been tracked
for at least one year and have been touched in the last year. The
number is simply the 12 month total minus the new listing number.
This hopefully discounts the coffee-break effect of the myriad 7DRLs
and lets us see how many "real" roguelikes are currently being updated
per year.
Year 6 12 % Total New Old
2004 16 25 27% 59 - -
2005 43 60 42% 103 +44 16
2006 32 47 36% 128 +25 22
2007 50 66 40% 163 +35 31
2008 54 70 36% 197 +34 36
2009 60 96 37% 262 +65 31
2010 107 120 35% 343 +81 39
2011 104 161 34% 469 +126 35
So where does eight years of data put us? We are doing very well for
roguelike creation - 10.5 new tracked roguelikes per month, a new
peak. It is tempting to dismiss this as a 7DRL effect, but the Old
column I think is correctly tracking the creation of larger projects.
The number of old projects is down to 35, matching our 30 a-year
theory. However, I supsect some of the new peak is a sampling
artifact as Michal Bielinski has properly captured some older
roguelikes missed on the earlier lists.
--
Jeff Lait
(POWDER: http://www.zincland.com/powder)
Ray
> It is now time for my eighth annual analysis of roguelike game
> development.
> The data for this comes from:
> http://thelist.roguelikedevelopment.org/
> which Michal Bielinski has been maintaining.
::SNIP::
> This tracks the number of roguelikes by last release date. The first
> column has a # for every roguelike released in the last month. I have
> omitted the last column which would have all the roguelikes over 42
> months old or without known release dates. There are now 123 such
> roguelikes being tracked.
In the interests of completeness, I decided to do the chart including
all games from that list for which the most recent release date is known.
I'm breaking it by quarter instead of by month, in part because I think
quarters show the trends more clearly. It may be a feeble reach for
sample sizes large enough for some statistical significance, if you like.
In this chart, each row is one quarter and each X represents one game
whose most recent update is in that quarter.
Emacs macros are fun.
Bear
1999 Q2 X
1999 Q3
1999 Q4
2000 Q1 X
2000 Q2 X
2000 Q3 X
2000 Q4
2001 Q1
2001 Q2
2001 Q3
2001 Q4
2002 Q1 XX
2002 Q2 XXX
2002 Q3
2002 Q4 XXX
2003 Q1 XX
2003 Q2 XXXX
2003 Q3 XX
2003 Q4 XXXXX
2004 Q1 XXX
2004 Q2 XX
2004 Q3 XXXX
2004 Q4 XXXX
2005 Q1 XXXXXXXXXX
2005 Q2 XXXXXX
2005 Q3 XXXXX
2005 Q4 X
2006 Q1 XXXXXXXXXXX
2006 Q2 XX
2006 Q3 XXX
2006 Q4 XXXXXXX
2007 Q1 XXXXXXXXXXXXXXX
2007 Q2 XXXXX
2007 Q3 XXXX
2007 Q4 XXXXXX
2008 Q1 XXXXXXXXXXXXXXXXX
2008 Q2 XXXXXXXXX
2008 Q3 XXXXXXXXXXXXXXXXXXXXXX
2008 Q4 XXXXXXXX
2009 Q1 XXXXXXXXXXXXXXXXXXXXXXXXXX
2009 Q2 XXXXXXXXXXXXXX
2009 Q3 XXX
2009 Q4 XXXXXXXXXXX
2010 Q1 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
2010 Q2 XXXXXXXXXXXXXXX
2010 Q3 XXXXXXXXXXXXXXXXXXXXXXXXXXXXX
2010 Q4 XXXXXXXXXXXXXXXXXXXXXXXXXXXX
2011 Q1 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
2011 Q2 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
Auric__
[snip]
> In this chart, each row is one quarter and each X represents one game
> whose most recent update is in that quarter.
>
> 1999 Q2 X
Which game is this one?
--
The infection has been removed. The soul of this machine has improved.
dormammu
you go. It seems to indicate the Roguelike genre as a whole is getting
more and more popular. Whether the popularity comes from the 7DRL (and
similar) or the 7DRL is a result of the increased popularity, I don't
know.
But the high failure rate is one of the reasons I don't plan to
officially announce the RL I'm working on (design stage currently, not
a single line of code has been written yet) until it's pretty close to
ready. If I never make it to that stage, no one will know.
Though is there a list of the names of which roguelikes have failed?
I'm always mildly curious to see what they're about and looking them
up would be interesting to me.
Ray
I dunno. I think we're discussing something orthogonal to failure
here.
These are games whose most recent release date was in some particular
timeframe. I think that's interesting as a matter of the history of
our hobby, but I don't get from there to "failure" in any reasonable way.
I mean, we can hardly say that Rogue "failed." But its most recent
release date is in 1986. Moria certainly didn't "fail." Its most
recent release date is in 1994. Nethack may have done many things,
but it did not fail. Its most recent release date is in 2003.
At what point did someone get it into their head that anything released
before a certain date was a failure, and why?
If you want to say something concrete about failure, I think you have to
actually look at the games themselves. My first qualification for
failure" is that the game is not available, or not in a playable state.
If a game is available and playable, then it is not a failure, no matter
when it was released.
Now, if you show me a game that is not available in a playable state,
whose developer/s are no longer working on it and don't plan to, then
I'm willing to call it a failure as a game. But a lot of even those
"failures" are learning experiences that contributed to the design of
later games.
Bear
Darren Grey
> dormammu wrote:
> > It's interesting to see that there are less failures the further back
> > you go. It seems to indicate the Roguelike genre as a whole is getting
> > more and more popular. Whether the popularity comes from the 7DRL (and
> > similar) or the 7DRL is a result of the increased popularity, I don't
> > know.
>
> > But the high failure rate is one of the reasons I don't plan to
> > officially announce the RL I'm working on (design stage currently, not
> > a single line of code has been written yet) until it's pretty close to
> > ready. If I never make it to that stage, no one will know.
>
> > Though is there a list of the names of which roguelikes have failed?
> > I'm always mildly curious to see what they're about and looking them
> > up would be interesting to me.
>
> I dunno. I think we're discussing something orthogonal to failure
> here.
>
> These are games whose most recent release date was in some particular
> timeframe. I think that's interesting as a matter of the history of
> our hobby, but I don't get from there to "failure" in any reasonable way.
Agreed, these annual posts should be more aptly entitled "release
rates", putting a more positive spin on things.
--
Darren Grey
Patric Mueller
> It is now time for my eighth annual analysis of roguelike game
> development. I shall first present some dubious statistics and then
> you shall complain that they don't accurately reflect roguelike
> development. Complainers may find some solace that my own roguelike,
> POWDER, has just fallen off the 6 month cut off to be an actively
> developing roguelike. The shame! The shame!
It happens to the best of us.
UnNetHack's last release was 9 months ago!
The iOS version of Powder (last update in January) doesn't count as it
is not in the datasources used for this statistics. That's what I call
a failure! ;-)
> So where does eight years of data put us? We are doing very well for
> roguelike creation - 10.5 new tracked roguelikes per month, a new
> peak. It is tempting to dismiss this as a 7DRL effect, but the Old
> column I think is correctly tracking the creation of larger projects.
> The number of old projects is down to 35, matching our 30 a-year
> theory. However, I supsect some of the new peak is a sampling
> artifact as Michal Bielinski has properly captured some older
> roguelikes missed on the earlier lists.
10.5 new RLs per month, there is clearly something I'm missing. I
should pay more attention to Roguebasin's news list.
Bye
Patric
--
NetHack-De: NetHack auf Deutsch - http://nethack-de.sf.net/
UnNetHack: http://apps.sf.net/trac/unnethack/
Michal Bielinski
> Complainers may find some solace that my own roguelike,
> POWDER, has just fallen off the 6 month cut off to be an actively
> developing roguelike. The shame! The shame!
My own creation is much further down the list so I will not be
the first person to chuck +1 stones in your general direction.
> I have
> omitted the last column which would have all the roguelikes over 42
> months old or without known release dates. There are now 123 such
> roguelikes being tracked.
This number is steadily increasing. Last year it was "only" 94. Some
of these have been lost permanently. I think Roguelike Graveyard
should be recreated albeit with slightly different goals. Many games
discontinued in middle of development are actually good for playing.
> The last column is Old. These are roguelikes that have been tracked
> for at least one year and have been touched in the last year.
Some of newer creations have their homepages linked with Github,
Sourceforge or other website with source code version control system.
Cataclysm has no actual releases but just updates to the code and no
proper version number. That way it is constantly close to the top
since I can count any commit as last development touch. If this trend
continues more and more regularly updated roguelikes will be counted.
> So where does eight years of data put us? We are doing very well for
> roguelike creation - 10.5 new tracked roguelikes per month, a new
> peak.
While roguelikes crop up out of the blue more often they also tend to
disappear much faster. This punches this number up.
MetaCollider was first presented on Sep 2008. Oct 2009 author takes
down SourceForge page and blanks RogueBasin entry. No signs of this
game since that time. Over a year of activity and its gone.
Legacy of a Warlord was released in Dec 2010. Last update was Mar 2011
and game page is no more. Author has a blog so this may be just on hold.
Sadly my experience tells such projects rarely get picked up.
HyperSilence first appeared on Nov 2010 and in Dec 2010 website had
disappeared.
> The number of old projects is down to 35, matching our 30 a-year
> theory. However, I supsect some of the new peak is a sampling
> artifact as Michal Bielinski has properly captured some older
> roguelikes missed on the earlier lists.
My efforts to fill missing roguelikes certainly had its effect. I find
it fruitful activity because old projects do get resurrected. For
example Dave resumed work on Kharne after a year of no updates and
Demonhunt 2.0 appeared in December last year after nine months.
--
Michal Bielinski