Topic: High Scores and Stats (Read 105271 times)

Kyzrati · « **on:** July 26, 2015, 10:48:11 PM »

Anyone who has opted to automatically upload their score sheet at the end of each run (set in the options menu) can check the high scores here.

High score updates only happen once per day. More details on the page itself.

(This is just a test system in preparation for leaderboards that will gradually expand to include more features, and maybe one day be automated.)

monsieurmarc · « **Reply #1 on:** August 18, 2015, 01:35:48 AM »

Very neat

Kyzrati · « **Reply #2 on:** September 01, 2015, 06:25:59 AM »

Update 150901: For probably a few days here I've paused the high score chart while I prepare a new one--bigger, better, all around awesomer. Alpha 3 will have a new test system during preparations for the upcoming tournament.

Adraius · « **Reply #3 on:** September 01, 2015, 09:06:22 PM »

Clarification? Have only the public updates paused, or has back-end score collection stopped? (if I get a killer score today, will it appear when the new chart goes up?)

Kyzrati · « **Reply #4 on:** September 01, 2015, 09:45:05 PM »

Sorry, I just meant that I'm temporarily not updating the web page (I've been doing that manually all along). The back-end is unchanged, and scores are still uploaded and received as always.

I will be using all previously collected Alpha 3 scores in the first new page when it's ready!

And, by the way, if for some reason someone doesn't have an internet connection when they end their game and you really really want your score on the board, you can also email me your scoresheet and I can just copy it over to the system. (This isn't so much an issue now with the temporary leaderboards, but could become more important during the tournament.)

Kyzrati · « **Reply #5 on:** September 02, 2015, 11:54:25 PM »

I've gone ahead and reinstated the original high score format (reset as of Alpha 3), since I've made some changes that will require releasing an Alpha 3c before it would be compatible with the latest and greatest leaderboards (currently WIP). So we'll continue using the old format for a little while longer.

Yes, Adraius, your win is up there

Kyzrati · « **Reply #6 on:** September 23, 2015, 02:05:32 AM »

Following the Alpha Challenge 2015, I've restarted the High Scores leaderboard and moved it to a slightly different location. It links to archived scores from previous lists, and will continue to receive updates as before (though once again not as frequently as during the tournament).

Kyzrati · « **Reply #7 on:** December 19, 2015, 06:55:56 AM »

New update to the High Scores page as of Alpha 5: It now also lists the furthest area reached by each player, to give more credit to speed builds. Note that where ties are concerned on the second table, the most recent record gets the higher spot.

As usual there will continue to be incremental improvements in the future.

Kyzrati · « **Reply #8 on:** March 25, 2016, 09:33:45 AM »

For the new release I reset the leaderboards again, and the first batch of participants is up now.

Something* tells me we won't be seeing as many ascensions with Alpha 7...

*Okay, Hunters.

Kyzrati · « **Reply #9 on:** April 28, 2016, 07:54:52 PM »

Reset the leaderboards again now that we have Alpha 8.

We've started off, unsurprisingly, with a higher ratio of cave-related deaths, including my own

Kyzrati · « **Reply #10 on:** June 09, 2016, 04:24:22 AM »

The Great Reset has occurred once again, this time for Alpha 9!

The final numbers for Alpha 8 stats show that in 258 reported runs, 17,165 robots were destroyed, among them 5,374 Grunts and 69 Behemoths.

In all, 22,238 parts were attached--3,421 power; 5,380 propulsion; 5,574 utilities; 7,863 weapons.

42,224 volleys were fired, averaging a maximum size of 2.74 weapons each. Out of 96,012 individual shots, 60,043 hit a target. 3.7% of shots were from a launcher.

1,546,152 points of damage was inflicted on targets, 47,748 (3.09%) of it self-inflicted via explosions.

1,357 targets were rammed, dealing 0.27% of all damage.

440 Haulers were intercepted (121 by Decker alone), 236 robots corrupted, and only 4 melted outright.

Target communications were jammed 9,839 times, 2,040 of those in a single run by Adraius.

261 Drones were launched, but only 7 ever recovered.

Average movement speed was 120%.

The average run visited 7 maps, while the single most expansive run, by Sherlockkat, saw 22.

An average 5 items were carried in inventory at any given time, with average inventory size peaking at 8.78.

2,679 machines were hacked an average 2.72 times each, with 44.2% of hacking attempts successful. 25.36% of hacks were entered manually; 27.44% of those were Trojans and brute force hacks. The most common hack was terminal records (lore).

21.28% of hacked machines were fully traced.

103 garrisons were disabled.

The most effective force of allies, led by Sherlockkat, inflicted 1,105 damage and made 14 kills.

Most dangerous ascension goes to Decker, who spent a quarter of an entire run at maximum alert (5).

Fastest win goes to zxc, who surfaced in only 1,108 turns, without destroying a single robot.

Only 11 pieces of alien tech were identified.

Kyzrati · « **Reply #11 on:** August 11, 2016, 12:26:21 AM »

Today the leaderboards were reset for Alpha 10.

Time for some stats from the last release!

Of 601 uploaded score sheets, we had a total 7 wins by zxc (3), Sherlockkat (3) and myself (1). Happylisk reported the highest score, 24,115. (Note: Old scores are always preserved in the archives linked at the bottom of the scores page.)

1,669 evolutions took place, the most popular slots being utilities (37.2%), followed by propulsion (25.2%), weapons (22.5%) and power (15.1%).

At each depth survived Cogminds had an average 75% core integrity remaining.

33,478 robots were destroyed, among them 11,640 Grunts, 7,594 Swarmers, 3,065 Sentries, 971 Programmers, 891 Hunters, and 81 Behemoths.

18 ARCs were destroyed before they could deploy, 1,723 annoying Recyclers were blasted, and desperate players took out 43 Compactors. Only 1 Executioner was eliminated.

Decker had the best combat robot kill streak, destroying 21 armed robots in rapid succession.

88,268 volleys were fired, averaging a maximum size of 2.69 weapons each. Out of 180,643 individual shots, 115,004 hit a target. 3.8% of shots were from a launcher.

2,748,369 points of damage was inflicted on targets, 85,007 (3.09%) of it self-inflicted via explosions.

2,979 targets were rammed, dealing 0.36% of all damage. 425 robots were corrupted, and only 6 melted outright.

191 robots were successfully hacked, 101 of them armed. 22 of those were assimilated, 19 overloaded, and 68 rebooted. Assimilated engineers tore down 8 walls, all in a single run by endlessblaze.

Enemy communications were jammed 11,219 times.

In all, 44,031 parts were attached--7,161 power; 11,231 propulsion; 10,456 utilities; 15,183 weapons. 19,867 of those attached parts were destroyed. Slot usage rate averaged 82.2%.

1,898,821 turns were passed, 15,924 (0.83%) while naked. 13 teleports were confirmed.

The average run visited 6.63 maps, while the single most expansive run, by Sherlockkat, saw 25.

540 derelict logs were recovered.

An average 5 items were carried in inventory at any given time, with average inventory size peaking at 8.56.

3,386 machines were disabled.

7,066 machines were hacked an average 2.72 times each, with 49.7% of hacking attempts successful. 24.1% of hacks were entered manually; 46.5% of those were Trojans and brute force hacks. The most common hack was terminal records (lore).

78 data cores were recovered, 27 of which were used.

11.7% of hacked machines were fully traced.

215 garrisons were disabled.

A total of 32 pieces of alien tech were identified.

Meta stats: 81% of players use fullscreen; 9.5% use hjkl for movement; 21% prefer ASCII; 27% don't touch the mouse. Smallcaps (the default) is the most popular font choice (52.4%), followed by Terminus (22.6%), and Cog (13.1%).

Kalkkis · « **Reply #12 on:** October 04, 2016, 11:44:04 AM »

Can I submit scores from before I had set them to be uploaded automatically? Here are few of mine.

Kyzrati · « **Reply #13 on:** October 04, 2016, 06:21:13 PM »

Sure! I'll add them to the batch. (You're doing pretty awesome already, it would seem

)

Kyzrati · « **Reply #14 on:** October 13, 2016, 01:44:10 AM »

The final Alpha 10 stat summary is in! (Scores have been reset, with old scores archived at the bottom.)

Of 915 uploaded score sheets, 618 had a high enough score (>=500) to be included in the stats. 95 unique players opted in to stat reporting, and we had a total 11 wins by zxc (5), Sherlockkat (2), Decker (2), Amphouse (1) and PleasingFungus (1). Decker blew everyone's score out of the water with his alert farming tactics, earning a 50,390 high. I can see scores continuing to rise as Cogminds have access to increasingly powerful components, plus numerous sources of bonus points via the new plot elements.

(note: some basic stats are not all that different from previous releases, and are omitted here--I'm just picking out things which might be interesting)

The highest number of bonus points earned in a single run was 27,640, by Sherlockkat, much of them from winning in only 3,131 turns.

920 garrison reinforcements were sent out, making up 13.2% of all squad dispatches. The most garrisons entered in a single run was 3, by ATM. 43 were visited in all, 248 were disabled from the outside, and 7 phase generators were destroyed during infiltrations.

36,321 robots were destroyed, 749 of them in the single-most destructive run by Decker. 26 Saboteurs were downed before they could detonate. Only 4 ARCs were destroyed before they could deploy. On the large end, 70 Behemoths were brought down, as well as 23 Compactors.

356 robots were corrupted; chain reactions were caused in 941 power sources.

48,175 parts were attached, 43% of which were lost. 34.7% of all parts attached were weapons, the most common category.

25 pieces of alien tech were recovered in all.

The highest speed achieved, by Sherlockkat, was 588%, or moving nearly 6 times per turn.

The most number of wheels used at once was 6, by Geralt.

Highest temperature reached was 2,446, by BoomBlip.

468 drones were launched, 33 recovered.

83 data cores were recovered, of which 29 were used. Of 22,780 machine hacks, 22.9% (5,211) were manual. 99 squads were recalled, including 54 investigations, 28 reinforcements, 15 exterminations, and 2 assaults. 2 Fabricators were overloaded. 40 robots and 155 parts were built. 322 parts were repaired.

54 traps were reprogrammed.

272 robots were rewired, 312 were successfully hacked (most commonly links and reboots, but there were 55 assimilations as well).

The average system corruption on death/winning was 2.6%, while the highest was 50%.

Meta stats:

83% of players use fullscreen
9.5% use hjkl for movement
26.3% prefer ASCII
30.5% don't touch the mouse
Smallcaps (the default) is the most popular font choice (64.2%), followed by Terminus (16.8%), and X11 (8.4%)
1080p is the most common resolution (59.0%), with 30.5% of players using something lower, and 10.5% higher

Kyzrati · « **Reply #15 on:** November 23, 2016, 01:55:59 AM »

Alpha 11 Stat Summary
(Scores have been reset, with old scores archived at the bottom.)

Of 1,563 uploaded score sheets, 1,097 had a high enough score (>=500) to be included in the stats. 113 unique players opted in to stat reporting, and we had a total 40 wins by zxc (19), Sherlockkat (9), Decker (9), Amphouse (1), Happylisk (1), and Timesplitter (1).

zxc achieved a new Cogmind all-time high score, 73,217 points for a speed run of only 764 turns (based on the new fast win scoring bonus formula). (He didn't record that one, but you can watch his 911-turn win , and read about the 764-turn win here.)

A few speed-running Cogminds are part of the reason for the higher than usual run count this alpha, but we also saw an influx of new players from the RPS article last month, right after the Alpha 11 update.

The average score was 4,947, though the median was much lower (2,762). Taking each player's best individual score, the average was 6,833 (median 4,547).

The most regions visited in a single run was 30, by zxc.

Congratulations to Ixoran for being the first to achieve the maximum possible speed, moving at 20 spaces per turn. Time itself practically stopped for this Cogmind.

There were 11,165 critical hits in all (will be interesting to see if this comes down following KI nerf) but only 1 robot was melted (<--we're definitely going to see that change in Alpha 12!).

Decker destroyed the most Behemoths in a single run: 5.

TrashT murdered 12 Zionites before succumbing to their wrath.

336 Assembled were destroyed.

Codemo fired the hottest volley of Alpha 11, generating +675 heat.

Some max individual run values, highest...
Hauler interceptions: 26 Robots corrupted: 32 Power chain reactions: 37 Communications jammed: 2,264 Parts field recycled: 114 Drone launches: 37 Traps triggered: 95 Garrisons disabled: 20 Robots built: 12 Parts built: 21 Parts repaired: 13 Robots hacked: 85 Robots overloaded: 45 Robots assimilated: 24 Manual hacks: 209 Trojans/brute force hacks: 76 Full trace events: 20

Most popular hacks by the number of successful results:
Zone Layout 1018 Record 3217 Part Schematic 900 Prototype ID Bank 729 Alert Level 700 Open Door 650 Locate Traps 565 Unreport Threat 487 Machine Controls 349 Level Access Points 346

Out of 38,500 machine hacks, 738 Trojans were installed (661 at terminals). 33.8% of hacks targeted machines other than terminals.

The top 5 lore collectors are
zzxc 83% Amphouse 62% Decker 60% Sherlockkat 48% Kalkkis 45%

The top 5 item collectors are
Decker 83% zzxc 77% Amphouse 71% Enno 70% Magi163 69%

209 pieces of alien tech were recovered in all, but by only 11 different players because yeah most of that stuff is later in the game.

The average system corruption on death/winning was 2.5%, while the highest was 50% (experienced by Chad).

Meta stats:

80.5% of players use fullscreen
14.1% use hjkl for movement
26.5% prefer ASCII
32.7% don't touch the mouse
Smallcaps (the default) is the most popular font choice (65.4%), followed by Terminus (17.6%), and Cog (6.2%)
1080p is the most common resolution (46.0%), with 46.0% of players using something lower, and 8.0% higher

Kyzrati · « **Reply #16 on:** November 27, 2016, 05:09:48 PM »

Leaderboard scores now link to score sheet stats! Note that if you've been to the page before, you might need to hit Ctrl-F5 to fully reload the new css before it'll appear normally for you (otherwise the colors could be off).

Spoiler: Example (click to show/hide)

Kyzrati · « **Reply #17 on:** January 20, 2017, 01:49:17 AM »

Alpha 12 Stat Summary
(Scores have been reset, with old scores archived at the bottom.)

Of 1,171 uploaded score sheets, 808 had a high enough score (>= 500) to be included in the stats. 125 unique players opted in to stat reporting, and we had a total 30 wins by Sherlockkat (15), lsend (4), Widmo (3), zxc (2), myself (2), ddarkray (1), Raato (1), Zaltroth (1), and BookofAeons (1).

zxc again broke the all-time high score record, with 90,820 points for basically owning the late-game and farming in an accidentally-too-easy Alpha 12. A single run spanning nearly 11 hours!

The average score was 5,319, though the median was much lower (2,862). Taking each player's best individual score, the average was 6,742 (median 3,742).

The most regions visited in a single run was 38, by Sherlockkat, a new all-time high.

Sherlockkat also disabled the most garrisons in a single run (19), followed closely by Widmo (18).

Widmo earns the title of most OP hacker, with a +200% bonus to hacking attack rolls. (We'll get a ton more interesting data like this with Alpha 13, now that more loadout capability-specific stats are recorded in score sheets!)

Regarding the damage type updates for Alpha 12--nerfed kinetics and buffed thermal, there were definitely some changes in what players were using. Compare the number of shots of each damage type:
Alpha: 11 12 KI 61.5% 51.9% TH 23.9% 30.7% EX 2.7% 3.8% EM 11.9% 13.6%

Maximums for total time spent on a given floor type:
(0b10) Materials Widmo 13,936 Factory Widmo 16,272 Research Kyzrati 4,727 Access Kyzrati 6,666 (other common areas) Mines Widmo 2,254 Storage Widmo 2,834 Garrison ddarkray 2,428 Lower Caves lord 5,008 Upper Caves ddarkray 3,485

Some max individual run values, highest...
Overflow Damage 1,091 (Amphouse) Heat Transferred 28,406 (Kyzrati) Parts Melted 54 (Kyzrati) Traps Reused 22 (ddarkray) Best Kill Streak 72 (Sherlockkat) Targets Rammed 60 (Toskin) Cave-ins Triggered 28 (dak) Self-inflicted Damage 1,429 (Widmo) Highest Corruption 54 (Kyzrati) Power Chain Reactions 37 (Widmo) Garrisons Disabled 19 (Sherlockkat) Robots Built 14 (277) Robots Rewired 25 (gammafunk) Robots Hacked 552 (zxc) Parts Built 31 (Sherlockkat) Traps Extracted 82 (Sherlockkat) Derelict Logs Recovered 14 (zxc) Exploration Rate (%) 78 (Widmo) Network Hubs Disabled 3 (Widmo) Actions Taken 46,572 (Widmo) Bonus Points 50,819 (Sherlockkat)

The top 5 lore collectors are
zxc 84% Sherlockkat 77% Widmo 70% Gressup 61% Amphouse 61%

The top 5 item collectors are
Widmo 81% zxc 73% Sherlockkat 69% Amphouse 68% Happylisk 67%

Meta stats:

89.6% of players use fullscreen (30% of those use borderless windowed mode)
16.8% use hjkl for movement (continues to rise with every release!)
20.8% prefer ASCII
36.8% don't touch the mouse
Smallcaps (the default) is the most popular font choice (70.4%), followed by Terminus (11.2%), and X11 (5.6%) (12.8% use some other font)
1080p is the most common resolution (44.8%), with 44.0% of players using something lower, and 11.2% higher

Kyzrati · « **Reply #18 on:** March 02, 2017, 09:29:41 PM »

Alpha 13 Stat Summary
(Scores have been reset, with old scores archived at the bottom.)

Of 968 uploaded score sheets, 696 had a high enough score (>= 500) to be included in the stats. 107 unique players opted in to stat reporting, and we had a total 15 wins by Sherlockkat (5), Shobalk (3), lsend (2), GJ (2), DDarkRay (2) and zxc (1).

Of course, while zxc only had a single win, it was his only run for the month and easily the highest scoring ever at a 26-hour "I own you all now" 276,549 points. (Now that Alpha 14 finally includes High Security, it's unlikely we'll ever see anything like that again!)

The average score was 5,668, a little higher than last time, with a higher median as well (3,048). Taking each player's best individual score, the average was 7,462 (median 4,932--definitely an upward trend).

Funniest manual seed goes to Senjai, who played a run with "cogmindhatesme", which ironically turned out to be the highest scoring of all his 14 A13 runs (9,063).

This time let's take a look at many of the new "Best States" recorded by score sheets as of Alpha 13:

Heat Dissipation 179 (Conner) Energy Capacity 2,530 (GJ) Matter Capacity 1,050 (Dekar) Sight Range 24 (GJ, Kyzrati, Sherlockkat, Shobalk) Robot Scan Range 30 (Sherlockkat, gammafunk, zxc) Terrain Scan Range 30 (nikola kolodziejczyk) Terrain Scan Density 2,410 (zxc) ECM Strength 8 (Sherlockkat) Cloak Strength 4 (MorlocK) Power Amplification 40 (Raine, lsend) Additional Mass Support 24 (Anonymous) EM Shielding 4 (Kyzrati, Sherlockkat) Armor Coverage (%) 99 (Conner) Resistance: KI 40 (gammafunk, Kyzrati, Shobalk, Xanthos) Resistance: TH 75 (Sherlockkat, Shobalk) Resistance: EX 75 (GJ) Resistance: EM 100 (Sherlockkat) Resistance: I 40 (gammafunk) Resistance: S 40 (gammafunk) Resistance: P 40 (gammafunk) Power Shielding 90 (eatnumber, Shobalk) Propulsion Shielding 90 (zxc) Point Defense Rating 240 (Amphouse, Sherlockkat, Xanthos) Thermal Conversion 180 (Raine) Weapon Cycling 50 (GJ) Melee Speed Boost 50 (Sherlockkat) Phase Shifting 30 (Sherlockkat) Evasion (%) 86 (Sherlockkat) Targeting Accuracy 40 (ParanoidSupernaut) Melee Accuracy 24 (buthix9) Launcher Accuracy 40 (Amphouse) Target Analysis 36 (Shobalk) Core Analysis 15 (Shobalk) Armor Integrity Analysis 100 (aaa, nikola kolodziejczyk, Mogwok) Recoil Reduction 10 (gammafunk) Overload Amplification 75 (ParanoidSupernaut) Overload Regulation 40 (Dekar) Stasis Canceling 3 (lsend) Offensive Hacking 237 (zxc) Defensive Hacking 54 (Shobalk)

Slight tweaks to Materials, and lots of extra beneficial encounters in the Mines, had the intended effect of reducing the number of early losses there. Deaths in the Mines especially dropped from 8.3% in Alpha 12 to 6.0%.

Among the new features that decreased the difficulty of some areas, 1,833 piles of Scrap were searched, the most in a single run by Shobalk (30).

Lore collection across all reporting players averages 11.0%; item collection stands at 20.5%.

Meta stats:

87.9% of players use fullscreen (32% of those use borderless windowed mode)
17.8% use hjkl for movement (still rising with every release!)
25.2% prefer ASCII
38.3% don't touch the mouse
Smallcaps (the default) is the most popular font choice (69.2%), followed by Terminus (15.0%), and X11/Dina (4.7% each) (6.4% use some other font)
1080p is the most common resolution (49.0%), with 39.8% of players using something lower, and 11.2% higher

Kyzrati · « **Reply #19 on:** May 10, 2017, 07:21:36 AM »

Alpha 14 Stat Summary
(Scores have been reset, with old scores archived at the bottom.)

Of 984 uploaded score sheets, 833 had a high enough score (>= 500) to be included in the stats. 122 unique players opted in to stat reporting, and we had a total 35 wins by GJ (10), MTF (6), lsend (5), zxc (5), Sherlockkat (4), Shobalk (3), may2ever (1), and Pimski (1).

The average score was 6,749, continuing to rise with every release. This is in part due to additional sources of bonus points, but overall player skill is increasing as well. (We'll probably see this rise a good bit in the next release now that there are bonus points for ally kills and at higher alert levels) Taking each player's best individual score, the average was 8,413, though the median (4,545) dropped slightly from Alpha 13.
By UI settings, based on best individual runs players using the mouse and/or tiles averaged around 8,000 points, while ASCII-and-keyboard purists earned an average 15,211.

31.9% of deaths were in Materials, 20.5% in Factory, 13.6% in some form of cave system, 8.5% in Storage, and 8.3% in Mines (17.2% were in other areas). In all, 25.5% of deaths were in the upper half of the world.

Altogether players achieved 2,765 evolutions. That's a lot of slots: 56.3% Utility, 22.6% Propulsion, 13.5% Weapons, and 8.1% Power.

High Security was triggered only 12 times, by GJ, MTF and Sherlockkat.

Alpha 14 was the first release to include easier difficulty modes! Most reporting players, however, stuck to the default roguelike mode:
Mode Runs Default 97.2% Easy 1.4% Easier 1.3%
Those playing easier modes tended to score higher--no surprise there. I'm sure we'll see those modes getting more use in the future on Steam. More players may even be using them now than we know, but are not the type who want to upload their scores, or have simply turned off uploading for the duration :P

Alpha 14 also added Cogmind's first two challenge modes, Scavenger and Unstable Evolution. There were 7 Scavenger runs, 1 by DDarkRay and the rest all by GJ, who won 3 of them. There were 10 Unstable Evolution runs, several by myself, Amphouse, and DDarkRay, while the remaining half were by GJ, again with the same three wins (because they were all double challenges :P).

For Beta 1, the latest iteration of the leaderboards has added a dedicated challenge mode for all those trying to score extra points by taking challenges. Check out the challenge thread for descriptions/details and set them in your /user/cogmind.cfg file to participate!

"Best States" record holders from Alpha 14:

Heat Dissipation 229 (Conner) Coolant Potential 240 (Anonymous) Energy Capacity 3,440 (GJ) Matter Capacity 1,200 (Pimski) Terrain Scan Range 30 (GJ) Terrain Scan Density 2,010 (Decker, GJ) Cloak Strength 4 (GJ) Power Amplification 60 (MTF) Additional Mass Support 32 (GJ, RandomString) EM Shielding 4 (GJ, Codemo, Pimski) Armor Coverage (%) 100 (GJ) Resistance: KI 55 (Phragmented, Zaltroth) Resistance: TH 60 (Phragmented) Resistance: EX 75 (GJ, MTF, Shobalk) Resistance: EM 100 (Sherlockkat) Core Shielding 40 (Conner, Kyzrati, MTF) Power Shielding 90 (Dekar, MTF, Nikola Kolodziejczyk) Propulsion Shielding 90 (Anonymous, WER6) Weapon Cycling 50 (GJ, may2ever) Melee Speed Boost 50 (MTF, Nikola Kolodziejczyk, Sherlockkat) Evasion (%) 93 (MTF) Targeting Accuracy 46 (GJ) Melee Accuracy 40 (Nikola Kolodziejczyk) Launcher Accuracy 40 (Conner, GJ, MTF) Target Analysis 32 (gammafunk) Core Analysis 15 (Codemo, GJ, Moredread) Armor Integrity Analysis 180 (GJ) Recoil Reduction 12 (Codemo) Particle Charging 100 (Moredread) Overload Regulation 60 (Pimski) Offensive Hacking 165 (MTF) Defensive Hacking 69 (MTF)

Lore collection across all reporting players averages 23.2% (more than double the Alpha 13 rate); item collection stands at 34.7% (also a significant increase). GJ had both the largest lore (97%) and gallery collection (99%), which took a big dive at the start of Beta 1 due to of all the new collectables.

Meta stats:

83.6% of players use fullscreen (27.5% of those use borderless windowed mode)
14.7% use hjkl for movement
24.5% prefer ASCII
32.0% don't touch the mouse
Smallcaps (the default) is the most popular font choice (75.4%), followed by Terminus (13.1%), and Cog (5.7%) (5.8% use some other font)
1080p is the most common resolution (45.9%), with 38.5% of players using something lower, and 15.6% higher

Kyzrati · « **Reply #20 on:** July 05, 2017, 04:00:56 AM »

Beta 1 Stat Summary
(Scores have been reset, with old scores archived at the bottom.)

Of 902 uploaded score sheets, 783 had a high enough score (>= 500) to be included in the stats. 125 unique players opted in to stat reporting, and we had a total 66 wins (a record high!) by Sherlockkat (13), zxc (13), lsend (9), GJ (8), MTF (7), Nikola Kolodziejczyk (6), Raine (3), DDarkRay (2), Palikka (2), Sabi Hakuhei (2), and Kyzrati (1). The higher number of wins can be attributed to three factors:
- Some players started using the easier difficulty modes
- Beta 1 added all seven different endings, sparking players to find how to collect them all
- And some of the regular players have been getting better :D

As predicted in the previous summary, the average score jumped to a new high again, reaching 8,035, though the best individual score average actually fell a bit to 7,238 as a larger number of new players joined for Beta (or returned for it after being away).

Beta 1 was only the second release since adding difficulty modes, so let's check out those stats again. Although most runs stick to the roguelike default, there's been a slight uptick in the number of players lowering the difficulty level, at least for some of their runs.
Mode Runs Default 95.0% Easier 2.8% Easiest 2.2%
9 different players played Easier mode; 5 played on Easiest.

Beta 1 added a new batch of challenge modes, of which there are now 8. It was also the first version with dedicated challenge leaderboards, occupied by a handful of players who'd already won the main game and wanted to try something a bit different. A total of 42 challenge modes were applied. The older Scavenger and Unstable Evolution challenges were only used twice each, while the most popular was clearly Pure Core (15). Super Gauntlet, undoubtedly Cogmind's most challenging mode, was played 9 times, while its tamer little brother Gauntlet was tried out just twice. All remaining three modes were tried out a bit as well, Devolution (4), Inhibited Evolution (2), and No Salvage (6). (You can read about the different challenges in their thread.)

"Best States" record holders (for select stats) from Beta 1:

Heat Dissipation 252 (GJ) Coolant Potential 360 (SpookUrDad) Energy Capacity 2,470 (Sherlockkat) Matter Capacity 1,800 (GJ) Power Amplification 180 (GJ) Additional Mass Support 25 (MTF, Palikka) Armor Coverage (%) 100 (RandomString) Resistance: KI 50 (Raine) Resistance: TH 60 (MTF, Sherlockkat, zxc) Resistance: EX 75 (DDarkRay, MTF, Sherlockkat, zxc) Resistance: EM 100 (GJ, Sherlockkat, zxc) Point Defense Rating 480 (Sherlockkat) Weapon Cycling 50 (GJ, Raine) Melee Speed Boost 50 (GJ, MTF, Noname1208, zxc) Evasion (%) 112 (zxc) Targeting Accuracy 60 (Palikka) Melee Accuracy 32 (Palikka, MTF) Launcher Accuracy 160 (Raine) Target Analysis 44 (Sherlockkat) Armor Integrity Analysis 100 (Antanst, Noname1208) Recoil Reduction 12 (GJ) Particle Charging 135 (zxc) Stasis Canceling 6 (Kyzrati) Offensive Hacking 194 (MTF) Defensive Hacking 78 (Palikka)

DDarkRay installed the most traps (39) in a single run (also setting a record high for traps triggered: 39 :P). Horse also installed quite a few (27).

Among the new mechanics, a fair number of players took advantage of tread crushing, but none came close to GJ's 152 robot-crushing run (also setting a ramming record of 711, since tread builds are immune to damage when ramming :P). BadToken (19), PyroL (17), and may2ever (15) kicked the most bots out of the way.

Melee builds have certainly matured quite a lot given the streak of buffs over many previous releases, and more players are relying on it as a support, or even primary, strategy. 14.9% of runs recorded more than 100 Cogmind melee attacks, and 7.0% had more than 200. There were 6,483 sneak attacks (max: 219 by MTF) and 2,326 follow-up attacks (max: 958 by zxc). Melee damage distribution:
Impact 5,402 (11.9%) Slashing 28,388 (62.5%) Piercing 11,658 (25.6%)

This time let's also analyze how players are choosing to traverse the world (or in some cases, "end up traversing" the world, since some people just wander or don't/can't check where they're heading first :D). For this I tallied the percentage of runs that entered a general branch, considering only those runs which passed through the depths from which the given area was accessible:
Min. Evolutions Visit% Area 2 42.9% Mines 3 18.4% Storage branch 3 27.2% Lower caves 5 31.5% Upper caves 6 43.8% Special caves 6 22.9% 0b10 Factory branches 8 44.5% Research branches

Plot-related stuff, described in cryptic terms to avoid spoilers:

Cogmind was imprinted in 14.7% of all runs; 23 players (2.9% of runs) decided not to imprint despite having the chance. 54.5% of winning runs were imprinted, with 5 winners (7.6%) choosing not to imprint.
4 runs interfaced with DC and went on to win, of 10 total runs choosing to do so.
34 winning Cogminds visited W, of 77 visits in all (44.2%).
13 wins were achieved with a reset core, out of 27 total resets (48.1%).

Beta 1 was the first opportunity to find alternative endings (and see all the different animations to conclude the story in different ways :D). 44 (two-thirds of) wins were of the easiest default type, "#0." Totals for other win types:
Win# Runs 1 4 2 8 3 4 4 0 5 1 6 5
No scores were uploaded for win #4, which isn't the most difficult but is a little harder to figure out.

Lore collection across all reporting players averages 23.3% (around the same as Beta 1, despite all the lore that was added then); item collection stands at 35.6% (also relatively unchanged). GJ had both the largest lore (96%) and gallery collection (99%), quite close to becoming the first person to have discovered everything.

Meta stats:

90.4% of players use fullscreen (29.2% of those use borderless windowed mode)
15.2% use hjkl for movement
17.6% prefer ASCII, a sudden new low!
29.6% don't touch the mouse
Smallcaps (the default) is the most popular font choice (64.8%), followed by Terminus (16.0%), and Cog (9.6%) (9.6% use some other font)
1080p is the most common resolution (44.8%), with 36.0% of players using something lower, and 19.2% higher

Kyzrati · « **Reply #21 on:** December 28, 2017, 05:16:02 PM »

Beta 3 Stat Summary
(Scores have been reset, with old scores archived at the bottom. I've also added section links since the boards can get quite lengthy.)

This time I've done something a little different than usual. Because we have a ton of new data, rather than just posting everything here in the forum thread I've made a bigger production of the stats with a bunch of graphs and analysis over on the blog. However, below I've also shared some player-specific stats from Beta 3 that you won't find in that post.

"Best States" record holders (for select stats) from Beta 3 (excludes anonymous players):

Heat Dissipation 265 (Happylisk) Coolant Potential 360 (squished) Energy Capacity 3,900 (Quelan) Matter Capacity 1,800 (Sherlockkat, Tron) Power Amplification 120 (Megumin) Additional Mass Support 36 (Tripdrake) Resistance: KI 40 (AlStar) Resistance: TH 75 (Fenrir, GJ, Malth, Valguris) Resistance: EX 75 (bugsniper, GJ) Resistance: EM 100 (GJ) Point Defense Rating 480 (Amphouse) Melee Speed Boost 50 (BigZombie, GJ, Highsight, nataya, PyroDawn, Quelan) Evasion (%) 109 (Quelan) Targeting Accuracy 64 (RBrandon) Melee Accuracy 32 (anonynamja, Quelan, Questking, Palikka) Launcher Accuracy 120 (Destravous) Recoil Reduction 12 (nataya) Particle Charging 50 (Dean) Force Boost 12 (Quelan) Salvage Targeting 15 (M.U.L.E.) Offensive Hacking 156 (GJ) Defensive Hacking 70 (Quelan)

Other personal single-run records that stand out among the others:

Legolas destroyed a mindboggling 1,597 robots. That was in Easiest Mode, but still...
Playing in regular mode, kiedra wasn't too far behind with 1,431 kills.
GJ used the most alien tech by far, taking all top three slots in that category.
GJ earned the most bonus points with 78,771.
Valguris has the honor of being the only player to die to system corruption, reaching 102% on the last floor. Not far behind was BigZombie, up to 93% corruption at death in Research
Palikka had more propulsion slots than anyone else, 13, during one of their wins.
Miruya used the most wheels at once: 6
Kyzrati had 8 weapon slots at once (heh I streamed that run, too--reached the end but it was suboptimal and I ended up losing :P)
GJ had the highest peak rating, at 248. Archivar and lsend tied for wins with the lowest peak rating: 77
GJ also earned the greatest number of points for follower kills: 11,374
Joking_Phantom and Raine tied for the most garrisons entered: 4
Highest win speed bonus goes to zxc, 41,152 points for winning in only 1,214 turns with 2,627 actions taken.
GJ destroyed 7 major NPCs.
Quelan had the longest kill streak of combat bots: 72
kiedra collected 61,401 matter.
kiedra also attached the most parts: 1,425 (59.8% utilities, 29.3% weapons)
zxc had the most consistent high slot usage, keeping slots 99% occupied through an entire winning run.
Zephyr survived 5,824 turns while naked.
BigZombie moved 32,623 spaces.
GJ was the only player to move 16 spaces per turn, just short of the fastest possible speed.
PyroDawn had the highest average speed: 410%
fqgglt suffered 2 propulsion burnouts.
Tron kicked 46 bots.
GJ crushed 22 bots.
Legolas corrupted 402 bots.
Jarchon melted 11 bots.
Zephyr triggered 161 cave-ins.
Raine had the greatest mass support: 310
Palikka ran around for the longest time while overweight by 10x.
GJ both carried the most items at once (52) and had the highest average carry rate (27).
fpsbossfight fired the hottest volley: 738
kiedra had 352 critical hit kills.
Kalkkis fired 213 overloaded shots.
BigZombie executed 1,427 melee attacks across a melee win.
Logrus had the most multiwielding followup melee attacks: 370
LuckyLuckLuc dealt 1,148 overflow damage.
Legolas reached a temperature of 2,116.
kiedra intercepted 50 Haulers.
Legolas triggered 151 power chain reactions.
squished jammed communications 19 times.
Lt Dan merge-repaired 19 parts.
GJ launched 68 drones
Ida recovered 24 drones.
Soar fabricated 36 traps.
BigZombie reprogrammed 52 traps.
Quelan triggered 65 traps.
Quelan also recovered the most data cores: 12 (but used only half of them)
BigZombie hacked 243 machines (782 hacks in all).
nink was fully traced 103 times.
PyroDawn installed 241 Trojans (78.4% on Terminals).
Quelan overloaded 3 Fabricators.
GJ built 20 robots.
BigZombie built 36 parts, and repaired 194.
Quelan rewired 53 robots.
BigZombie hacked 121 robots (among them 65 combat bots).
GJ had 22 allies hacked, but successfully repelled 3 other attempts.
GJ had the most allies (255) and the largest group at once (35).
GJ triggered high security 5 times (and still won!). 359 squads were dispatched that run (300 of them assaults).
Pimski visited the most regions: 38
DDarkray spent the most time in Garrisons (13,467 turns).
GJ spent the most time in Waste (1,777 turns).

Kyzrati · « **Reply #22 on:** January 28, 2018, 06:25:43 PM »

Beta 4 Stat Summary
(Leaderboards have been reset.)

Time for another stat summary! Note it's just this forum post, whereas I won't be doing such a complete analysis as I did for the Beta 3 stats--that was a special case since we suddenly had so many new players from the Steam release.

8,793 score sheets were uploaded over previous Beta's four weeks, 7,097 (80.7%) of those with a high enough score (>=500) to be included in the stats. 4,923 (69.4%) of the remaining runs were anonymous, however, so the stats will focus on the 2,174 runs by named players. There were 653 unique players in all, 432 (66.2%) of which were non-anonymous.

Not unlike Cogmind's alpha launch back in 2015 there was a huge surge of players for Beta 3, and the run count dropped off over the next month or two, though still propped up a bit by the new surge of players during the Winter Sale. This Beta was also pretty short, only taking four weeks to release, so that also contributed to its lower run count.

There were 67 wins by Pimski (16), lsend (6), MojoPoop (6), Dullahan (5), Valguris (4), Airus (3), fpsbossfight (3), zxc (3), Destravous (2), Owlsley (2), Raine (2), and all of the following players with one win each: atomfullerene, bugsniper, DDarkray, Demdiran, GridBugBear, happylisk, Joshua, Kyzrati, Nikola Kolodziejczyk, nuzcraft, Terminus, Wad67, wilk.

After GJ reigned supreme for a number of releases, in Beta 4 one of our newer players Valguris topped the leaderboards with a score of 103,974, a mostly hover-based 10-hour combat run through 30 different maps. fpsbossfight wasn't far behind (96,177) with his 28-map 5-hour heavy combat run. The other three Top 5 wins were all players competing over pacifist speed runs, Pimski escaping in only 888 turns, followed by zxc (1,003) and lsend (1,123).

As expected, the average score has started to rebound after dropping significantly with the surge of new players. It's now up 19.1% to 5,364, although still a fair bit lower than pre-Steam levels.

The difficulty mode spread remains little changed from before:
Mode Players Default 93.1% Easier 3.7% Easiest 3.2%

Of the seven endings, only the final and most difficult two had no recorded wins, because the few players capable of achieving these endings were mostly on break for Beta 4.

To serve as a comparison for the next release in which imprinting has been significantly modified, let's record the current state of things:

Cogmind was imprinted in 7.5% of runs, and in most cases (74% of the time) where players had the chance they chose to imprint.
37.5% of winning runs were imprinted, with 7 winners (10.7%) choosing not to imprint.
Players called for dispatches 368 times (average 2.27 per imprinted run), 78.5% reinforcements (21.5% resupply).
Intel was retrieved 138 times (average 0.85 per imprinted run).
Both dispatch and intel data shows a relatively even distribution in terms of type, mainly because not everything is always available and Cogmind is forced to choose from less desirable options over the course of a run. The low average values can be attributed to the portion of players who either succumb to the dangers of the return trip, or are destroyed before long in the mid game.

Including anonymous score sheets, in the four weeks of Beta 4 players...

evolved 16,538 times
explored 41,971 maps
recovered 6,614 Derelict logs
searched 10,202 piles of scrap
kicked 4,777 robots
crushed 989 robots
fired 1,068,440 volleys
inflicted 33,209,009 damage
destroyed 411,924 robots
absorbed 476,396 damage via shields
entered 367 Garrisons
disabled 279 Relays
destroyed 40 Network Hubs
carried out 10,423 sneak attacks
intercepted 10,076 Haulers
created 14,737,924 salvage
attached 477,428 parts
used 486 pieces of alien tech
hacked 71,823 machines
entered 31,354 hacks manually
used 13,151 unauthorized hacks
hacked 3,395 robots
built 500 robots
fabricated 1,392 parts
took 17,829,131 actions
passed 22,848,511 turns
earned 26,693,647 points

"Best States" record holders (for select stats) from Beta 4 (excludes anonymous players):

Heat Dissipation 287 (Terminus) Coolant Potential 320 (TurtleHat) Energy Generation 122 (Valguris) Energy Capacity 4,580 (josiah) Matter Stores 1,696 (fpsbossfight) Matter Capacity 1,800 (fpsbossfight) Terrain Scan Range 30 (GridBugBear, Pimski, s0lid, Terminus) Terrain Scan Density 2,110 (sim) ECM Strength 12 (Eb) Power Amplification 90 (lsend) Additional Mass Support 30 (Barbarossa) Corruption Prevention 40 (nuzcraft) Resistance: TH 75 (Airus) Point Defense Rating 320 (josiah) Weapon Cycling 50 (Deemzul, fpsbossfight, GridBugBear, s0lid) Melee Speed Boost 50 (AlmondRoll, Draxis, Raine) Evasion (%) 108 (Pimski) Targeting Accuracy 56 (Lt Dan) Melee Accuracy 38 (Terminus) Launcher Accuracy 70 (s0lid) Recoil Reduction 15 (josiah) Force Boost 5 (Destravous, Pimski) Salvage Targeting 9 (Terminus) Offensive Hacking 148 (Terminus) Defensive Hacking 73 (Terminus)

Other personal single-run records that stand out among the others:

Valguris used the most alien tech, 20 pieces in all.
Pimski earned the most bonus points in a single run: 70,111
Terminus reached 79% corruption before dying in Research.
RustyCougarMama and wilk both evolved 5 power slots. More power to them.
Pimski had more propulsion slots than anyone else, 15, during a speed run win.
bobduude and NomnNomn both had the most weapon slots: 7
Valguris had the highest peak rating, at 231.
lsend, Pimski, Raine, and zxc all earned the maximum pacifist bonus (10,869).
Pimski earned the maximum pacifist bonus on 4 separate runs.
fpsbossfight earned 17,417 bonus points for high alert combat kills.
Joshua earned 5,127 bonus points for follower combat kills.
MojoPoop entered the most garrisons: 4
Valguris destroyed 4 Network Hubs on all of three runs.
Sherlockkat lost the most points to friendly fire: -6,897
fpsbossfight destroyed 29 different classes of robot.
fpsbossfight felled 7 Behemoths.
Valguris obliterated the Fortress on default difficulty.
Valguris killed 4 major NPCs.
Pimski and Valguris had the longest kill streak of combat bots: 36
fpsbossfight collected 33,698 matter.
Terminus lost the most parts: 311
Pimski survived 1,091 turns while naked.
Pimski had the highest average speed: 384%
Destravous made 639 overloaded moves.
GridBugBear rammed 72 bots.
sim kicked 35 bots.
Krinok and Raine each crushed 8 bots.
Valguris had the heaviest build, 314, and greatest mass support: 600
fpsbossfight and Valguris tied for largest inventory capacity, 38, though only fpsbossfight carried up to 34 spare parts.
Sherlockkat carried an average of 19 parts throughout a run.
Terminus took 54,185 damage.
Terminus fired 1,978 volleys.
Valguris absorbed 8,006 damage with shields.
mindreader fired the hottest volley: 770
fpsbossfight fired 5,604 shots.
josiah fired 461 launcher shots.
Kyzrati set the gunslinging record with 104 secondary targets (totally won that run).
bugsniper fired 36 overloaded shots.
Valguris executed 625 melee attacks.
zorser sneak attacked 130 robots.
AlmondRoll had the most multiwielding followup melee attacks: 112
Valguris dealt 1,417 overflow damage.
Marbot reached a temperature of 2,223.
Valguris purged the most corruption: 61
Valguris intercepted 49 Haulers.
fpsbossfight corrupted 217 robots.
josiah triggered 61 power chain reactions.
neutral jammed communications 21 times.
bugsniper launched 28 drones
Valguris recovered 10 drones.
DDarkray reprogrammed 47 traps.
Draxis triggered 43 traps.
Valguris hacked 621 machines, performed 280 manual hacks, and installed 142 Trojans.
A Legendary Sword overloaded 4 Fabricators.
Vissk built 10 robots.
Terminus built 18 parts.
TurtleHat repaired 10 parts.
Valguris rewired 32 robots.
Airus and Joshua tied for the most allies: 93, though Pimski also had the largest group at once (29).
Valguris triggered high security 4 times in two separate runs, and still won both.
Pimski (once again) visited the most regions: 36
MojoPoop spent the most time in Garrisons (3,212 turns).
khiro spent the most time in Waste (1,526 turns).

Meta stats:

93.6% of players use fullscreen (29.0% of those use borderless windowed mode)
9.0% use hjkl for movement
11.0% prefer ASCII
17.3% don't touch the mouse
0.9% of players have rebound at least one command
Smallcaps (the default) is the most popular font choice (64.5%), followed by Terminus (16.2%), and X11 (8.1%) (11.2% use some other font)
1080p is the most common resolution (53.5%), with 29.2% of players using something lower, and 17.3% higher

Vi keys, ASCII, and keyboard mode are all decently rebounding (doubling on average) following Cogmind's Steam release as the more hardcore players stick with it long term.

Kyzrati · « **Reply #23 on:** May 13, 2018, 10:34:08 PM »

Beta 5 Stat Summary
(Leaderboards have been reset.)

With the detour to work on the 7DRL and huge achievements release, we've accumulated quite a few Beta 5 score sheets in the meantime!

12,535 score sheets were uploaded from late January to early this month, 10,261 (81.9%) with a score high enough (>=500) to be included in the stats. 6,027 (58.7%) of those were anonymous so stats below focus on the 4,234 runs by named players except where indicated. There were 569 unique players in Beta 5.

There were 206 wins by Pimski (29), lsend (16), Joshua (15), Leebears (15), Owlsley (13), GJ (12), Valguris (12), Malth (10), fpsbossfight (9), Laida (8), GridBugBear (7), Terminus (7), PlasticHeart (6), Lt. Dan (5), Mojopoop (5), Sherlockkat (5), Ape (3), bugsniper (3), FluffyCat (3), wilk (3), ario (2), MTF (2), zxc (2), and all of the following players with one win each: b_sen, Bruh, Doctor, Dullahan, GDFKTT, Jazzer, mindreader, MSK Impostr, neutral, nuzcraft, RNGesus, Vissk, Xavius.

The fairly high win rate came as a result of quite a few players becoming experienced enough to reliably win (:D), while several others shared their first ever Cogmind win during Beta 5.

By type, the basic w0 was the most common, accounting for more than half (126, 61.2%) of all wins. The other more easily discovered win types (w1/w2/w3) had more winners (11/41/16, respectively) than the very challenging and elusive w4/w5/w6 (2/3/7).

Four players broke the 100k score barrier this release, all of them ending with mostly hover combat builds and fighting into the extended end game. Valguris led with a steep 166,008, followed by Terminus (132,541) who didn't quite make it out in one piece, GJ (119,799), and fpsbossfight (104,848). A lot of the other 50k+ leaderboard runs were also combat-focused this time around, more than usual considering we have so many long-time players normally dedicated to flight strategies.

Average score rose again, adding 31.9% since Beta 4 to reach 7,076. Lots of newer players who joined us after the Steam release are clearly getting better.

Once again we have about the same distribution among difficulty modes (which I'll stop reporting on in the future):
Mode Players Default 94.1% Easier 2.2% Easiest 3.7%

Last time I recorded a number of imprinting-related stats with the intent of comparing them to Beta 5 since that was one area that saw significant adjustments and new features (see release notes for more info). Here is an update on what changed, or didn't:

Cogmind was imprinted in 9.1% of runs, and in most cases (60.6% of the time) where players had the chance they chose to imprint. This was not too different from last time, although the 60% value is down from 74%, partially because 20% of the time players instead opted to do something in Zion which precludes imprinting (a spoilery strategy that became more popular during Beta 5).
28.6% of winning runs were imprinted, with 52 winners (46.8%) choosing not to imprint. Overall this was much less imprinting than before, most likely due to that changes making it less ideal for hackers to imprint. It remains a powerful strategy for combat or hybrid builds (as originally intended), but is no longer a free power boost for other types of builds.
Players called for dispatches 2,356 times (average 6.2 per imprinted run), 70.7% reinforcements (29.3% resupply).
Despite imprinting less, those who did so made far more use of it! While the reinforcement:resupply ratio was little changed, the average number of dispatches per imprinted run tripled, likely because one of the major new attractions in Beta 5 was the new category of, um... friends.
Intel was retrieved 3,056 times (average 8.0 per imprinted run). The increase here was even greater than the reinforcements, up by a factor of eight! As of Beta 5 imprinting makes regular Terminal hacking impossible, so players needed to rely on extra intel to fill the infowar gap.

"Best States" record holders (for select stats) from Beta 5:

Heat Dissipation 292 (MTF) Coolant Potential 580 (Ann) Energy Generation 170 (Valguris) Energy Capacity 5,740 (Valguris) Matter Stores 1,232 (Vogel) Terrain Scan Density 2,210 (PlasticHeart, Pimski) ECM Strength 6 (Avendir, Dullahan, Horse) Power Amplification 160 (Valguris) Additional Mass Support 36 (GJ) Corruption Prevention 24 (Doctor, Grunt, Molten Bear) Armor Coverage 2,850 (GJ) Point Defense Rating 800 (GJ) Evasion (%) 111 (Pimski) Targeting Accuracy 64 (Pimski) Melee Accuracy 66 (Pimski) Launcher Accuracy 120 (ario) Recoil Reduction 14 (mindreader) Particle Charging 50 (Malth, Terminus) Force Boost 8 (AlmondRoll) Salvage Targeting 6 (Leebears) Offensive Hacking 300 (GJ) Defensive Hacking 300 (GJ)

Other personal single-run records that stand out among the others:

GJ used the most alien tech, 19 pieces in all.
Valguris earned the most bonus points in a single run: 95,109
Sherlockkat reached 100% corruption and died in in the extended game.
Pimski and mindreader had more propulsion slots than anyone else, 20.
Kyzrati reached the end with the most weapon slots: 8
Jazzer and Leebears used the most wheels at once, 5.
GJ had the highest peak rating, at 275.
Pimski earned the highest pacifist bonus (12,004).
Valguris earned 43,069 bonus points for high alert combat kills.
MojoPoop earned 7,784 bonus points for follower combat kills.
Pimski entered the most garrisons: 23
Valguris and GridBugBear destroyed 4 Network Hubs.
Pimski earned the highest win speed bonus: 65,530
GJ lost the most points to friendly fire: -21,519
Valguris destroyed 37 different classes of robot.
kiedra eliminated 127 Programmers.
GJ felled 9 Behemoths.
fpsbossfight and Valguris executed 13 Executioners.
fpsbossfight downed 5 Alpha 7s.
Jazzer and Valguris murdered 48 Zionites.
Valguris killed 12 major NPCs.
GJ obliterated nearly 200 robots in a single turn.
kiedra collected 44,347 matter.
Leebears attached the most parts: 1,137
Pimski won while attaching the least number of parts: 33.
Malth and Pimski tied for highest average speed: 569%
GJ made 3,707 overloaded moves.
GJ rammed 152 bots.
GJ kicked 149 bots.
Leebears crushed 15 bots.
Valguris had the heaviest build at 350.
GJ and Valguris both carried an average of 27 spare parts throughout a run.
Terminus absorbed 21,452 damage with shields.
GJ fired 3,407 volleys.
ApolliniaD fired the largest volley: 8 weapons.
Alexbot fired the hottest volley: 740
kiedra fired 8,858 shots.
GJ fired 412 launcher shots.
kiedra set the gunslinging record with 96 secondary targets.
Taraityg chained together the most gunsling targets at once: 3
kiedra killed the most bots with critical hits: 322
may2ever fired 114 overloaded shots.
Raine executed 1,944 melee attacks.
Raine sneak attacked 234 robots.
Raine had the most multiwielding followup melee attacks: 339
Valguris dealt 963 overflow damage.
Horse reached a temperature of 1,732.
kiedra purged the most corruption: 101
kiedra intercepted 70 Haulers.
Valguris corrupted 402 robots.
Leebears triggered 59 power chain reactions.
Vogel jammed communications 33 times.
bugsniper was simultaneously tracked by 210 hostiles.
kiedra made 113 tactical retreats.
GJ launched 54 drones
ApolliniaD recovered 30 drones.
Valguris built 240 traps.
Joshua reprogrammed 26 traps.
Valguris hacked 218 machines and performed 476 manual hacks.
fpsbossfight installed 176 Trojans.
PolymathAspirant overloaded 5 Fabricators.
Sabazin built 32 robots.
Pimski built 22 parts.
redd repaired 15 parts.
Pimski collected 21 robot analyses.
Terminus rewired 34 robots.
quetzalcoatl hacked 70 robots.
PlasticHeart assimilated 36 robots.
GJ had 204 allies across an entire run.
MTF had 44 allies at once.
GJ spent 42% of a run at Alert Level 5.
Pimski (once again) visited the most regions: 85 (this time via garrison looping in Access)
Lt Dan spent the most time in Garrisons (3,669 turns).
Valguris spent the most time in Waste (1,837 turns).

Meta stats:

91.4% of players use fullscreen (23.8% of those use borderless windowed mode)
6.2% use hjkl for movement
11.8% prefer ASCII
19.2% don't touch the mouse
1.1% of players have rebound at least one command
Smallcaps (the default) is the most popular font choice (67.1%), followed by Terminus (17.1%), and X11 (8.1%) (7.7% use some other font)
1080p is the most common resolution (54.8%), with 32.5% of players using something lower, and 12.7% higher

Vi key usage fell significantly as a wider variety of players from outside the genre discovered Cogmind this year. ASCII use has remained fairly steady, however, and there are slightly more keyboard mode players now.

Note that due to new regulations in the EU, as of Beta 6 anonymous scores are no longer uploaded at all, and score uploading is opt-in only (as it was prior to the Steam release), so we'll have a little less data to go on than before, but no matter, we've gotten a glimpse of the anonymous situation and there's not a whole lot of interesting or unexpected information there anyway. We're mostly interested in the stats provided by regular players contributing to the leaderboards :)

Despite the drop in total data, more part-time players will probably be active for Beta 6 now that we have achievements to go for! That'll be the big thing to examine next time.

Kyzrati · « **Reply #24 on:** September 11, 2018, 12:27:35 AM »

Beta 6 Stat Summary
(Leaderboards have been reset.)

Beta 6 was a big release since it added the 256-achievement system, and it also took us quite a while to hit Beta 7 due to the huge new robot hacking system, so that means 6 was out there for quite a while to give us... more data!

However, remember that due to the GDPR changes in May, score uploads were switched to opt-in only so we no longer get any anonymous data, bringing us back to the way it was before last year prior to Steam. That's fine with me, although it means what we're seeing here now only represents a subset of the player base. As for the portion of players we do have here, based on old data I'd say it's at least half.

Against that background, let's dive in...

[For a while now I've kinda had a formula going for these stat analysis posts, but I won't be following that so closely this time.]

In the four months (!) of Beta 6, 5,619 score sheets were uploaded, of which 4,811 (85.6%) had a score high enough (>= 500) to be included in the stats. 529 unique named players submitted at least one run, averaging 9.1 qualified runs per player.

Altogether Beta 6 saw a stunning 462 wins (an 8.2% win rate!) by Valguris (81), Mhorre (38), Malthusis (30), lsend (27), Raine (26), Terminus (22), wilk (22), Pimski (19), zxc (19), GJ (16), PlasticHeart (15), Main.Cuck (12), GridBugBear (11), MojoPoop (11), Decinym (9), PhenomPhear (9), GOR (8), Zyalin (8), Joshua (7), Palikka (7), Ape (4), Ben'n'Dan (4), FluffyCat (4), Airus (3), Amphouse (3), bugsniper (3), Kyzrati (3), Rumbl3 (3), Alexbot (2), DDarkray (2), fpsbossfight (2), James Vagabond (2), Maugrift (2), Owlsley (2), Quelan (2), Sabazin (2), SNACKS (2), and all of the following players with one win each: b_sen, Benjamin, Donnie, Finestep, Leebears, LuckyLuckLuc, MajesticThrust, Misery, Mitaro, nameofteam, nuzcraft, Phantom, RBrandon, RNGesus, Tapuck, TheMulletron, VapourLock, Xifax, and Zinc. (Note that most of these players hang out on our Discord, so stop by if you're looking for advice :D)

Even factoring in the community's increasing collective skill, those are some pretty high numbers, especially those topping the list. One clear factor is the popularity of turn-based speedrunning as an alternative way to play Cogmind, and a lot of the speedrunners are quite good so it's not really a matter of whether they'll win, just how long it'll take. Assuming wins below 3,000 turns are probably speedruns, they account for half (49.4%!) of all wins. This makes sense because speedrun wins are pretty quick to knock out compared to regular runs.

I got in a speedrun win myself, unfortunately just outside of 2,000 turns (though only my second try :D). I remember when I used to be able to reliably hang out in or near the top 10 on the leaderboards, but now my usual runs aren't even in the top 30 anymore! There are more players, and more players are getting better. Beta 6 leaderboards can be found in the archives, including all Beta 6 score sheets linked at the bottom.

Achievements being the big thing in Beta 6, we should definitely take a look at those.

Except for the few which were bugged and only just recently fixed for Beta 7, at this point every achievement has been earned.
although no single player has earned all achievements yet. The highest achievement rates so far:
Player Achievement Rate GJ 89% Valguris 79% Raine 73% PhenomPhear 72% Terminus 71%

Note that we can't take a close look at specific achievement numbers because unfortunately Steam does not display accurate rates. The data you see on their site is rounded, and Cogmind has so many challenging achievements compared to the number of frequent players that a bunch of achievement numbers simply round to zero. They've all been earned by at least someone though.

Among the rarest achievements are some of the win types, since there are so many different ways to win and not many of the associated achievements can be earned simultaneously. Really a good number of achievements in general are mutually exclusive, so truly covering the full range is a matter of doing enough runs while aiming for these different achievements--using different play styles, taking different routes through the world, engaging with different challenges and plot points...

That's not to say you can't earn a lot of achievements in a single run. The top achievement-earning runs:
Player Most Achievements in a Single Run Valguris 124 zxc 95 Pimski 93 lsend 76 Airus 74

124 is nearly half of Cogmind's achievements in a single run! Of course, that was as 15-hour combat run that hit a crazy amount of content...

In total players earned 135,575 achievements during Beta 6, averaging 28.2 per run.

We also had a bunch of score sheet entries added specifically as part of the achievements update to help record information required to award them, so we could check out some of those as well...

195 Cogminds were scanned by Researchers, which can be a dangerous thing to have happen, but some players either don't mind or even (!) do it intentionally. 60 winning runs were scanned.
0.8% of matter was collected remotely, the same ratio among both wins and non-winning runs.
28,856 unidentified prototypes were attached, about 6 per run. Terminus attached the most in a single run (120).
917,801 spaces were dug, 69,045 (7.5%) via melee. Certainly the majority of terrain was destroyed by AOE weapons, but melee digging must be explicit so that high a percentage is telling of the value of melee digging as an effective strategy.
14,174 Recyclers were shooed. McSi wins the award for being the most annoyed, having shooed 47 Recyclers in a single run.
29 machines were destroyed in a single turn by PhenomPhear. Boom.
22,027 unidentified exits were taken, an average of 4.58 per run. GJ took the most in a single run: 24

Meta stats:

91.5% of players use fullscreen (25.8% of those use borderless windowed mode)
6.6% use hjkl for movement
11.7% prefer ASCII
23.3% don't touch the mouse
0.9% of players have rebinded at least one command
Smallcaps (the default) is the most popular font choice (63.5%), followed by Terminus (18.7%), and X11 (8.7%) (9.1% use some other font)
1080p is the most common resolution (55.0%), with 30.1% of players using something lower, and 14.9% higher

There were almost no changes to the meta stats since Beta 5, aside from a noticeable increase in the number of keyboard players and the portion of those selecting the Terminus font somewhat rose against Smallcaps.

This time around we also got a new meta stat to look at: Tactical HUD use. 73.2% of players have activated that feature. Anecdotally some players prefer having somewhat fewer numbers and less clutter on the interface, although a majority of those players using the regular HUD probably either didn't play much yet or haven't noticed it in the options. If you aren't using the Tactical HUD yet, you'll most likely want to do that once you're familiar with most of the mechanics since it gives extra useful information (but can be confusing and distracting when you first start out, hence it's off by default).

In terms of purely aesthetic preferences, only 2.1% of players use a render filter, unchanged from Beta 5 (but I forgot to cover that last time, when that feature was first introduced). The few who use it are generally turning on low-contrast mode.

Next time we'll get to take a look at the robot hacking system introduced for Beta 7, among other new stats added to the score sheet.

Grid Sage Forums

News:

Author Topic: High Scores and Stats (Read 105271 times)