Page 2 of 3

Re: Map-Twitter

PostPosted: 10.01.2017, 18:30
by Seraph07


a quick comparison between the recently played maps from today and from the day I started to change the mappool (21.12.2016).
As you can see the additional 111 maps, which where added at the 31.12.2016, make a difference in the top 25 score.

But how does it effects the mappool in total?
Games of the last 30 days
today*: 6752
before**: 6110

Here is the maplist from both dates compared to each other.
Played maps_10.01.2017 & 21.12.2016.xlsx
(24.63 KiB) Downloaded 463 times


Re: Map-Twitter

PostPosted: 13.01.2017, 12:11
by Seraph07
Updated the Highscore with antique informations from 2014.


You can find the scource here:
Recent games

Total games from 03.10.2014: 13324

Please remember that the hard drop in the highscore between 2016 and 2017 affected only the top 25 maps. Total games where risen up ~10%.

Re: Map-Twitter

PostPosted: 16.01.2017, 13:58
by Seraph07
Updated Maplist
All maps_16.01.2017.xlsx
(24.74 KiB) Downloaded 478 times

All maps_16.01.2017.csv
(2.9 KiB) Downloaded 451 times

Do we add today more maps than in the past?

Year Maps added

Maps added per year.png
Maps added per year.png (6.66 KiB) Viewed 20650 times

Maps added per year.csv
proof of the data
(25.24 KiB) Downloaded 473 times

Update 24.01.2017

Maps Fun 246
Maps Tournament 25
Maps Testing 64

Maps Total 335

Re: Map-Twitter

PostPosted: 31.01.2017, 14:04
by Seraph07

I updated the minimum requirement: a map needs 15 games per month to stay in the mappool. Maps in Testing need 10 games to get an upgrade into FUN.

Here is a list of the maps which droped out of the mappool

133 maps droped out.
127 Fun
6 Tournament
104 [16x16]
27 [32x16]
2 [32x32]

Comments to some maps
- Manta 8 games today, 03.10.2014 135 games, Top 23 map
- Weave 6 games today, 10.12.2016 95 games, Top 9 map
- Stones 11 games, 9 games from one player in one day
- Camouflage 49 games, 29 games from one player, 29 in a row!
- Redworld 15 games, a once loved map near at the edge

Maps which lie back in the alphabet seems to get fewer games.
Maps like "3Towers" will now only get added as "ThreeTowers" or "Towers3".

Solo survivor players will have the biggest impact on this statistic.

They can play the most games in a day because there is no waitingtime. It does not matter if a game has four player or just one. All vs all games have a smaller impact in comparison.


All good maps are missing?!?
    The unloved maps droped out, you can still play them but there wont be any skill or xp anymore.

    Maps that droped out will join the mappool if they get 10 games in a month. They will get a new chance in the next maprotation.

How can I play the "Testmaps"?
    Just type the mapname in the "Load Map" field while you create a game. My tip is to play some survivor games, this will push the recent games the most.

Why do I always get +0 or +1 xp points for solo survivor?
    New maps need 10 games for the survivor mode until the xp award is normal. It is deactivated because it would be to easy to push your xp if a new record at round 10 would give you +10xp.
    Established maps will get you normal xp, new maps need some time.

(22.02 KiB) Downloaded 474 times

Statistic: =sum(recent games)

Re: Map-Twitter

PostPosted: 31.01.2017, 19:10
by Seraph07
57 brand new maps!

35x [16x16]
17x [32x16]
5x [32x32]

new maps_31.01.2017.csv
(3.5 KiB) Downloaded 464 times

Re: Map-Twitter

PostPosted: 02.02.2017, 11:03
by Seraph07
Game distribution from new maps.

What map do you wanna play? You open the maplist and click on the image. Many will click on the first half of the maplist right?
I tryed to proof this on some recently added maps. Some have 0 games (Wunderbar) and some 15 (Anemone).

First I listed all maps in an alphabetical row, splitted them in half, mapnames from A - M and N - Z and added the recently played games together.

sum games.png
sum games.png (5.33 KiB) Viewed 20590 times

As we can see the hypothesis works for all maps and 16x16 maps in particular.
But 32x16 and 32x32 are conspicuous.

So how many maps do we have in each category and how is the distribution?
distribution mapnames.png
distribution mapnames.png (4.98 KiB) Viewed 20590 times

How does the distribution affects the sum? I divided the sum of games through the sum of maps.
games & maps divided.png
games & maps divided.png (5.22 KiB) Viewed 20590 times

16x16 maps tend to be played more if they are placed at the start of the alphabet.
The distribution of 32x16 and 32x32 maps is not equal on the alphabet but it is no problem.
32x16 maps are played equal regardless of the alphabet.
32x32 maps tend to have a name in the first half of the alphabet.
32x16 and 32x32 maps are played intended which works against the hypothesis.


remember that this is only for the maps added at the 31.01.2017 and only for 57 maps!

My data
game distribution from new maps.xlsx
(19.41 KiB) Downloaded 470 times

For all maps
sum games_all.png
all maps = Fun & Tournament
sum games_all.png (5.75 KiB) Viewed 20588 times

distribution mapnames_total.png
distribution mapnames_total.png (5.33 KiB) Viewed 20588 times

games & maps divided_total.png
games & maps divided_total.png (5.44 KiB) Viewed 20588 times

game distribution_total.xlsx
(37.88 KiB) Downloaded 465 times


How is the distribution from the maps that droped out of the mappool?

Re: Map-Twitter

PostPosted: 05.03.2017, 01:56
by Seraph07
Maps that droped out of the mappool.

New maps

Re: Map-Twitter

PostPosted: 09.03.2017, 22:41
by Seraph07
Egoistic and I looked at the current mappool and found some bugged games in the highscore.

Resetted Highscore Bounty Hunter for following maps.

Mineshaft ... 4hlqtfb7oi

Sober ... i8jmpzueyg

Special ... 6v0js790sd

Re: Map-Twitter

PostPosted: 14.03.2017, 15:23
by Seraph07
Some lists from the maps ingame for Knappe8.
Saved as .csv (MS-DOS)
(3.69 KiB) Downloaded 452 times

(3.67 KiB) Downloaded 456 times

(531 Bytes) Downloaded 468 times

Re: Map-Twitter

PostPosted: 15.03.2017, 17:45
by Seraph07
Big Survivor Update

The survivor values for all official maps (Fun & Tournament) changed to lower values → higher scores and new survivor records are near to catch!
That was the maininformation. The following text is about the how and what changed.

Endless gameplay
The survivor values for this maps was to low, so they have now higher values.
The Big Combo Highscore on this maps is resetted.

Easy ... h93aih2z83
Medium ... xfh88nxj2z

Old survivor values
Easy: 125
Medium: 130
Hard: 135

New survivor values
Easy: 131
Medium: 137
Hard: 143

Easy ... hn7kn00eji

Old survivor values
Easy: 120
Medium: 125
Hard: 130

New survivor values
Easy: 123
Medium: 128
Hard: 133

Easy ... 7e7py96i95

Old survivor values
Easy: 117
Medium: 121
Hard: 125

New survivor values
Easy: 120
Medium: 125
Hard: 130

Easy ... 0qg5htl814

Old survivor values
Easy: 125
Medium: 130
Hard: 135

New survivor values
Easy: 128
Medium: 134
Hard: 140

Easy ... 9lrlpp9jgv

Old survivor values
Easy: 131
Medium: 138
Hard: 145

New survivor values
Easy: 135
Medium: 142
Hard: 149

Some maps are near to perfect
I reduced the survivor values -1% on this maps.

Easy: 118
Medium: 114
Hard: 107

Easy: 122
Medium: 95
Hard: 0

Easy: 89
Medium: 111
Hard: 102

Maps with noticeable lower survivor values
(easy to get rank 1 here)

All maps that are not listed above are reduced by 2 or 3%.

Re: Map-Twitter

PostPosted: 15.03.2017, 18:22
by Seraph07
  • What are "surivor values"?
    The KI for the Big Combo mode has its own income but if it would be the same as normal players (100%) it would be easy to outdeff the KI and it could not defeat you. The KI needs a higher income on long and easy to deffending maps (120-140%).
  • The distribution between Easy, Medium and Hard is not the same, why?
    Small changes on the income will have a huge effect on short Maps like Speedvector (90%, 92%, 94%) but the effect is not so big on long and easy to deff maps like Quadcore (135%, 142%, 149%). So we need a higher distance to separates the wheat from the chaff. The difference on Speedvector is 2% and on Quadcore 7%.
  • What does -X% mean?
    The survivor values for Big Combo Hard changed about X%.
    Old survivor values for Bubbles
    Easy: 117
    Medium: 121
    Hard: 125

    The survivor values changed -4% (as you can see in the list above). The KI loses 4% on its income for Big Combo Hard. The other values orientate on the Hard value (list of the distribution).
    New survivor values
    Easy: 113
    Medium: 117
    Hard: 121

    That should be the normal case for most maps. -4% equals -4% on all Big Combo modes. But there are exceptions...
    As we saw above the distance between Hard and Medium is not always the same (Speedvector 2, Quadcore 7). Because of this situation does a reduction of the survivor values on Hard not always decrease the Easy values the same way.
    Easy Medium Hard
    107 110 113
    106 110 114
    107 111 115
    108 112 116
    Reducing the Hard values from 115 to 113 does not effect the Easy values at all. Only Hard and Medium changed.

    This can be a problem! If Easy and Hard drift to far away from each other it will be hard to change the Easy values if they produce endless gameplay. Big Combo Hard will be way to high as we can see on some of the longer maps. Further informations will be following.

Re: Map-Twitter

PostPosted: 16.03.2017, 10:40
by Seraph07
Seraph07 wrote:I played a few survivor games recently and it is true that the values are very high on most maps.

There will be changes this week.

Reducing the survivor values for hard 3-5% and using the old distribution for easy.
Easy = Medium - ((Hard - Medium) x 2)

This way every map and especially Big Combo Easy will be much easier.

Quote from comes from the suggestion to lower the survivor values.

That was my original plan. But as I started changing the survivor values I reached a decision to stay with the current distribution of survivor values. The survivor values for Big Combo Hard are reduced and thereby the Easy and Medium values too.
The current distributions has some flaws and changing it back to a larger distance between Hard and Easy wont solve the problem, it would only make it bigger.

We had five maps with registered endless gameplay, so the survivor values for Easy had to be raised.

Old survivor values
Easy: 125 <- endless
Medium: 130 <- endless
Hard: 135

The new survivor values for Big Combo Easy had to be higher than 130.

This is the current distribution of the survivor values for Atomic. The difference between the modes is 6%, between Easy and Hard 12%.

Current list
Easy: 125 -> 131
Medium: 130 -> 137
Hard: 135 -> 143

The second list would double the distance between Easy and Medium (Easy = Medium - ((Hard - Medium) x 2)). That is the distribution I announced before (14.03.2017).

Second list
Easy: 125 -> 131
Medium: 130 -> 143
Hard: 135 -> 149

I have only values from the first list last to 140% on hard (110 134 140), so I will use this distribution for the example (Easy to Medium 24% and Medium to Hard 6%).

First list
Easy: 125 -> 131
Medium: 130 -> 155
Hard: 135 -> 161

The two old distributions would raise the Hard values way to high and would lead to short games even on maps like Atomic.

If there are still questions feel free to ask.

Have a nice day and creep on.

Re: Map-Twitter

PostPosted: 19.03.2017, 22:32
by Seraph07
MJochen, DerSpetzI and Earthquake are Maptester now.
They will help to find the best survivor values for all maps and help testing the new maps.

Re: Map-Twitter

PostPosted: 20.03.2017, 16:56
by Seraph07
I need your help on following maps

Blocked fields yes or no?

Upgrading a deffspot, 1 or 2 rounds?

New pathdesign required.

New pathdesig required

New pathdesign required.

Re: Map-Twitter

PostPosted: 20.03.2017, 18:11
by Seraph07
Next month will 18 new maps join the mappool.

At the current state 27 maps will drop out.


Some little changes in the maphighscore

Games last 30 days

(50.75 KiB) Downloaded 466 times

Re: Map-Twitter

PostPosted: 21.03.2017, 20:31
by Seraph07
Bugged game on Bloodlines ... r9n9nuu7rz

Bounty Hunter is resetted.

Re: Map-Twitter

PostPosted: 25.03.2017, 09:46
by Seraph07
A quick reminder that if the survivor values are to low they will raise up and the highscore is resetted.
The demand to lower the values will resett some maps. I cannot say how many maps it will hit but there will be many (Asturcon, Corrodor and Zauberwald atm).

The maptester who test the survivor values will lose their highscores too. It is a great passion from the maptester to test the survivor values and have to earn their rank again after the update.

Thank you guys for testing the current values and help to improve the CreepTD survivor.

Re: Map-Twitter

PostPosted: 27.03.2017, 12:20
by Seraph07
Some statistic on how many games we have.

Recently played games (games in the last 30 days).
27.03.2017 --> 5048

Re: Map-Twitter

PostPosted: 27.03.2017, 16:08
by Seraph07

Re: Map-Twitter

PostPosted: 01.04.2017, 15:40
by Seraph07
New maps



Recent games
Fun 4413
Tournament 934
Testing 236

Total 5583