1v1 and 2v2 matchmaker pools for March

@BIG-BENNIS-MAGIC 2? - it is quite blue, but a really refreshing map in terms of layout and gameplay

I do not like maps in which the entire area is red - the eyes get very tired very much

Thanks to the ever amazing @arma473, I have been able to make a spreadsheet with the frequency that ladder matches on particular maps are aborted. I consider a game aborted if it ends within 90 seconds (initially I was doing 3 minutes, but it seems there are some openers such as first bomber, intie rush killing transports, etc. that will potentially end the game within 3 minutes). The data are composed of ladder matches going back to March 1st 2020 and are broken down by ladder rating.

@Valki while I’m not saying you’re wrong, I will say that your anecdotal experience about people quitting frequently on Badlands simply does not agree with the empirical data. 8 - Badlands_v4 has a 1.92% abort rate at <300 and a 2.12% abort rate at 300-800. For reference, the median abort rate for the <300 bracket is ambush the enemy with a 2.91% abort rate and is Vulcan's Reach with 4.44% for at 300-800. In fact, out of the 88 maps that have been played by ladder players at 300-800, only 8 maps have a lower abort rate than Badlands.

@Brutus5000 it seems your experience does match the data. For the 300-800 bracket, outside of Moonlight Mesas with a 20.95% abort rate, the next 4 aborted maps are all 20x20 maps (Seton's Clutch: 10.11%, Roanoke Abyss: 10.24%, Seraphim Glaciers: 11.13%, Crossfire Canal - FAF version: 15.00%). These are very high abort rates and should be taken into account.

Other interesting stats:

• 62/105 maps in the 1800+ bracket have a 0.00% abort rate.
• The highest abort rate maps are in the 1800+ bracket with Adaptive Kusoge at 22.72% and Broken_Vows at 22.22% (small sample size).
• The average* abort rate in the <300 bracket is 3.40%.
• The average* abort rate in the 300-800 bracket is 5.04%.
• The average* abort rate in the 800-1300 bracket is 4.19%.
• The average* abort rate in the 1300-1800 bracket is 3.03%.
• The average* abort rate in the 1800+ bracket is 2.47%.

*average calculated by weighting every map equally.

Overall, the higher the rating, the lower the abort rate. Perhaps the better you get, the less picky you are with maps because you are more comfortable playing a variety of maps (with a few notable exceptions). Or, perhaps the willingness to play on a greater variety of maps makes you a better player. Not sure which direction the causation arrow points. Also not really sure why 300-800 is the pickiest bracket by a decent margin.

Going forward for future ladder pools, I will also take abort data into consideration as an additional criterion when making the map pools.

@MarcSpector if you’d like to go through the trouble of going through the entire ladder map pool and marking each map by color I could take color of map into consideration.

300-800 is probably the pickiest bracket because of a single prolific player. The Moonlight Mesas abort rate is largely because he did that 104 times in a row. Which was almost 8% of all ladder matches on that map for an entire month of FAF.

In my bracket (300-800) you often have the problem that the games tend to take very long even on smaller maps.
Like both players turtle at least a little bit and its hard to get a killing blow (unsure if you have actually enough units, lack of ideas, lack of skill). So many games I played I knew I had already won after 15 minutes still took 40-45 minutes.
I could assume the time here massively scales up with the map size. Many naval maps 20x20 take over an hour and even don't make much kind of "amazing" gameplay.

@archsimkat said in 1v1 and 2v2 matchmaker pools for March:

@Valki while I’m not saying you’re wrong, I will say that your anecdotal experience about people quitting frequently on Badlands simply does not agree with the empirical data. 8 - Badlands_v4 has a 1.92% abort rate at <300 and a 2.12% abort rate at 300-800. For reference, the median abort rate for the <300 bracket is ambush the enemy with a 2.91% abort rate and is Vulcan's Reach with 4.44% for at 300-800. In fact, out of the 88 maps that have been played by ladder players at 300-800, only 8 maps have a lower abort rate than Badlands.

I have no trouble believing I just had bad luck - I play too little to award any statistical significance to my experiences.

However, sometimes I try to look up the replay to take note of who left on me, but often I do not find a replay. Do your abort rate statistics register games that are too short to generate a replay?

Really happy you are taking a look at this though, I have little time to play and spending 2-3 matchmaking rounds to get an "abort" can waste my play opportunity of that moment.

I think it not hard to understand the data. Low skillers suck at multitasking and therefore refuse to play 20x20. Fixed by rating specific pools. Then there are maps that offer to much to do, and its obvious that its too much. Like kusoge, you need to drop plateaus, grab reclaim, secure expansions, and not get gunship sniped. Its impossible to do all tasks, and the constant frustration of knowing you arnt doing these makes this map and all other maps like it an utter brain fuck to play.

Im not saying that good map design has no reclaim and a finite number of expansions and plateaus to drop, but i have a strong feeling that this might be the case. Look at farms map, no reclaim bullshit, you actually need to think about unit movement and investment more than anything. Whoohooo, new game???

Honestly reclaim isnt doing the game a favour. Its a nice idea, yeah, but it shifts gameplay from units to macro and well, if I want that I play sim city.

Maps like Kusoge are very difficult to play and that's why they only appear in 1800+ brackets. They force players to compromise and prioritize, you have only time to do X amount of things and you need to choose well if you want to succeed, that's why they can be frustrating to play but also offer the greatest challenge and playing well on them gives you a lot of satisfaction, knowing that only few players are able to play well on such map.

@Valki I don't know what games are "too short" to generate a replay. A lot of the replays are literally 3 seconds long, which should mean that the person leaves before they even get control of their ACU.

Are you referring to people hitting the cancel button during the loading screen? If that happens, no match ever starts, so there is no replay and no record of it goes into the vault.

However, players are not supposed to know which map/opponent they have, until the match starts. Your own client won't tell you who you have been matched against until the match starts. In order to use the cancel button to avoid a certain map or a certain opponent, you would need another person helping you. They would have to use their FAF client to look up who your map/opponent is, and tell you, quickly enough that you could hit the cancel button. I wouldn't worry about that happening.

@arma473 these are games where I do enter into the game, but "x has been defeated" is the first thing I see, immediately on entering.

I have never seen a replay of such a game when I checked.

I assumed in these cases my opponent enters the game first and also leaves it before I enter.