Afterlife Rated 1.18 Statistics

[dwarftough]

Hi everyone! I've got a couple of requests to gather stats for Afterlife Rated 1.18. As there are 5000+ games already, we can have a decent sample size to judge what's going on (although, as ratings were reset and it's just 5000 raw games, that wouldn't compare to Afterlife 1.16 year's stats).

Links to previous topics with Afterlife stats:

• viewtopic.php?t=55715
• The final year's stats for Afterlife 1.16

This is stats for Afterlife Rated 1.18, played till November 3, 2024. It uses Afterlife Era and standard XP mod. Only reported games that lasted at least 7 turns were considered.

As ratings are fresh, I will only give three categories: 1500+, 1600+ and 1650+. Matchups for each category are sorted from the most equal to the most onesided. "A vs B: 50.01%" means faction A wins against faction B 50.01% of games.

1500+ players, 1833 games

Spoiler:

Knalgan Alliance vs. Northerners : 50.81% ( 124 games)
Loyalists vs. Undead : 52.21% ( 113 games)
Undead vs. Northerners : 54.46% ( 112 games)
Undead vs. Knalgan Alliance : 55.97% ( 134 games)
Knalgan Alliance vs. Rebels : 56.92% ( 130 games)
Drakes vs. Loyalists : 57.76% ( 116 games)
Loyalists vs. Northerners : 59.31% ( 145 games)
Rebels vs. Loyalists : 59.59% ( 146 games)
Rebels vs. Drakes : 60.0% ( 115 games)
Rebels vs. Undead : 62.64% ( 91 games)
Knalgan Alliance vs. Loyalists : 63.85% ( 130 games)
Rebels vs. Northerners : 66.39% ( 119 games)
Knalgan Alliance vs. Drakes : 67.44% ( 129 games)
Undead vs. Drakes : 69.23% ( 104 games)
Drakes vs. Northerners : 69.6% ( 125 games)


Player 1 won 887 (48.39%)
Player 2 won 946 (51.61%)

1600+ players, 584 games

Spoiler:

Loyalists vs. Undead : 52.5% ( 40 games)
Knalgan Alliance vs. Rebels : 54.29% ( 35 games)
Undead vs. Northerners : 54.76% ( 42 games)
Drakes vs. Loyalists : 58.82% ( 34 games)
Rebels vs. Loyalists : 59.52% ( 42 games)
Undead vs. Knalgan Alliance : 62.0% ( 50 games)
Knalgan Alliance vs. Drakes : 63.41% ( 41 games)
Northerners vs. Knalgan Alliance : 63.83% ( 47 games)
Loyalists vs. Northerners : 65.22% ( 46 games)
Rebels vs. Undead : 70.59% ( 34 games)
Rebels vs. Drakes : 72.92% ( 48 games)
Drakes vs. Northerners : 73.33% ( 30 games)
Knalgan Alliance vs. Loyalists : 75.61% ( 41 games)
Rebels vs. Northerners : 82.86% ( 35 games)
Undead vs. Drakes : 84.21% ( 19 games)

Player 1 won 283 (48.46%)
Player 2 won 301 (51.54%)

1650+ players, 424 games (Note: this is extremely unreliable, as some matchups played as rarely as just 10 games here. Although it may be a sign by itself)

Spoiler:

Drakes vs. Loyalists : 51.85% ( 27 games)
Knalgan Alliance vs. Rebels : 53.85% ( 26 games)
Loyalists vs. Undead : 56.25% ( 32 games)
Rebels vs. Loyalists : 60.0% ( 35 games)
Northerners vs. Knalgan Alliance : 60.61% ( 33 games)
Undead vs. Knalgan Alliance : 61.11% ( 36 games)
Undead vs. Northerners : 62.07% ( 29 games)
Loyalists vs. Northerners : 64.52% ( 31 games)
Knalgan Alliance vs. Drakes : 66.67% ( 27 games)
Rebels vs. Undead : 70.37% ( 27 games)
Drakes vs. Northerners : 73.68% ( 19 games)
Rebels vs. Drakes : 75.68% ( 37 games)
Knalgan Alliance vs. Loyalists : 75.86% ( 29 games)
Rebels vs. Northerners : 77.78% ( 27 games)
Undead vs. Drakes : 88.89% ( 9 games)

Player 1 won 208 (49.06%)
Player 2 won 216 (50.94%)

[Cremember]

ty for posting these stats!
few ideas i got from these stats
1. rebels are way too strong 4 favorable matchups all 60%+
idea to fix: nerf shaman/sorceress +20% xp cost and increase the cost of shamans to 17g
2. northerners are clearly the weakest 4 matchups 40%-
idea to fix: -25%xp cost for orcish chief(making it similar to elvish and drake leadership)1-2 and 2-3 both.
idea 2: shamaness is completely useless give it magic attack and +1 range dmg
3. ud vs drakes seems to be the most overwhelming one. the reason u have so little stats is because some of us have agreed to not even play the matchup
idea to fix: give all drake units +10 cold resis so -40% for fighter and clasher and -30 for burner
4. dwarves seem like a harder fix basically they do well in matchups where thunders are effective(loyal and drakes)its the single best unit to get a strick
idea to fix: +15%xp cost for thunders both 1-2 and 2-3 and -15%xp cost for guards 1-2 and2-3(this can help in the orc matchup)

these are just some stuff off the top of my head. As we can see in the higher lvs the gaps are even bigger the favorable matcups just wins too often.
our goal should be to get everything as close to 50% as posiible, got a long way to go now we just focus on the 60%+ ones. we are in need of some balancing work done assp!!

[Pagawan]

Perhaps the balances required call for a comprehensive review of the underlying factions. From abilities , weapon specials, resistances, defences, damage types, hp, and mps. This is the only way to get wholly balanced factions that gives equal chances no matter the matching.

[dwarftough]

3. ud vs drakes seems to be the most overwhelming one. the reason u have so little stats is because some of us have agreed to not even play the matchup

To be honest, I'm surprised that it's now firmly believed to be so one-sided. It clearly felt like a very dangerous matchup for Drake, but in 1.16's years stats on the top levels it's like just 56-60% winrate of Undead and normal amount of games, so more or less bearable and with mutual chances

[matto]

3. ud vs drakes seems to be the most overwhelming one. the reason u have so little stats is because some of us have agreed to not even play the matchup

To be honest, I'm surprised that it's now firmly believed to be so one-sided. It clearly felt like a very dangerous matchup for Drake, but in 1.16's years stats on the top levels it's like just 56-60% winrate of Undead and normal amount of games, so more or less bearable and with mutual chances

What ever made AI behave different is the cause, i might be crazy tho?

[dwarftough]

What ever made AI behave different is the cause, i might be crazy tho?

Well, Afterlife doesn't apply any AI settings, so it uses the defaul AI provided by the game. It obviously could change from version to version (I think it may change even between minor versions), although I'm not sure if it really changes in 1.16 -> 1.18 (considering, afaik, that AI maintainer is, up to my knowledge, either a vacancy or less active).

Still, I wonder, what exactly do you think AI started doing differently in 1.18?

[lhybrideur]

Was the behavior making AI go to heal itself to villages/healer/rest healing not improved? I remember reading something like that, but I do know neither if it was about default AI nor if it was added in 1.18?

[matto]

What ever made AI behave different is the cause, i might be crazy tho?

Well, Afterlife doesn't apply any AI settings, so it uses the defaul AI provided by the game. It obviously could change from version to version (I think it may change even between minor versions), although I'm not sure if it really changes in 1.16 -> 1.18 (considering, afaik, that AI maintainer is, up to my knowledge, either a vacancy or less active).

Still, I wonder, what exactly do you think AI started doing differently in 1.18?

I don't think i am able to describe what is different, but i give it a try

This one isn't Afterlife related but AI got weirdly attached to Spam Units (in World Conquest II), since it got into mainline.



Meanwhile for Afterlife 1.16 to 1.18 now AI changed its targets.

1.16 targets were pretty straight forward, in the early waves AI ranged units would suicide on the player ranged units(and viceversa), in 1.18 they attack melee units, slowing the killing process, a great exemple is an ulf on the standard afterlife map, if an assassin, an archer and 4 to 5 grunts can reach the ulf on mountain tile near the village (leaving 2 hex to attack him), will be attacked by the archer and assassin, instead of having multiple grunts suicide like in 1.16. {THIS tends to make units with both ranged and melee shiny more, it allows the player to kill AI units easly}

(1.18) When AI units move after spawn, they now tend to blob.

1.16 AI stutters when spawning got more common in 1.18 (when the unit doesnt in the first turn of spawn)

1.18 AI tends to miss 99.99% ctk on lower level units if there is a possibility of "killing" a lvl2 or higher, which sometimes results in a bad choice and makes AI more unpredictable.

Recently i tested out if a charger would prefer to kill a goblin or a lvl2/1, and it choose the lvl2, i guess it took into account the value?


If something gets in mind or i find the perfect replay exemples i will send them, hopefully XD

(EDIT!)

To be honest, I'm surprised that it's now firmly believed to be so one-sided. It clearly felt like a very dangerous matchup for Drake, but in 1.16's years stats on the top levels it's like just 56-60% winrate of Undead and normal amount of games, so more or less bearable and with mutual chances

Btw Undeads always had a "100%" winratio with Zombies vs Drakes. Talking strictly about top players, never saw Shigune lose a game of Ud vs Drakes


---------------------

Attached a replay of me playing Undead vs Dwarves, Undead in 1.16 would win (without zombie) because Dwarvish Fighters suicided on Ghost, in 1.18 they avoid ghost, unless they are built with 15dmg x6strikes, leading footpads poachers or anything with ranged to attack them, slowly stacking units until your ghosts get weak enough and die, here i played with zombies because of that. Btw we can find very cool exemple of what i was saying about

1.18 AI tends to miss 99.99% ctk on lower level units if there is a possibility of "killing" a lvl2 or higher, which sometimes results in a bad choice and makes AI more unpredictable.

My ghosts didnt go to kill a 9 Hp footpad and an Ulf on flat with 15~20 hp, instead decided to "suicide" on the lvl3 Dwarvish Lord with 30+- hp on 50% dodge with 5 strikes.

[Pagawan]

Can the number of forest/ woods tiles be reduced and replaced with more neutral tiles like sand, mushrooms, and cave tiles? I noticed one of the reason elves dominate is there is usually a lot of forest/woods tiles especially located in key areas.

[Bad_tactics]

Game 5306 I disconnected and returned in less than 5 minutes only to find opponent reported it as their victory and left.

[dwarftough]

Game 5306 I disconnected and returned in less than 5 minutes only to find opponent reported it as their victory and left.

I've cancelled that game. This topic is for discussion on stats and balance, for contested games please use the topic in the Tournaments section: viewtopic.php?t=58614

[Bad_tactics]

Game 5306 I disconnected and returned in less than 5 minutes only to find opponent reported it as their victory and left.

I've cancelled that game. This topic is for discussion on stats and balance, for contested games please use the topic in the Tournaments section: viewtopic.php?t=58614

Will do in case of other similar issues. Thanks.

[Alleria]

Hi There,
Just reflecting on the ideas of Cremember - In some cases, Crem's suggestions are a bit radical, but generally I agree. Summarizing my thoughts below

Original quote:
1. rebels are way too strong 4 favorable matchups all 60%+
idea to fix: nerf shaman/sorceress +20% xp cost and increase the cost of shamans to 17g

My thougths:
I believe +10 xp cost (or something similar that gives a round number as a result) and + 1 gold increase is enough as first step - let's see, how it works and we can make further adjustments in the future, if stats are still unbalanced.

Original quote:
"2. northerners are clearly the weakest 4 matchups 40%-
idea to fix: -25%xp cost for orcish chief(making it similar to elvish and drake leadership)1-2 and 2-3 both.
idea 2: shamaness is completely useless give it magic attack and +1 range dmg"

My thoughts:
The purpose of orcish chiefs are having 1 - 2 of them in an army and no more. Given that, XP buff is not that necessary, however, we can try to lower it by 10% or something like that. Of course, when you try a strategy of spamming 4-5 orcish chiefs in an army (which I see often), you will face difficulties leveling them up
It would be essential to give shamaness magical attack - so that Northerners could have a chance against elves and drakes as well.
+ 1 damage doesn't matter much, since we can increase it via XP mod, so I would suggest to give magical attack first - then (if statistics are still bad for orcs), we can increase base damage by 1.

Original quote:
3. ud vs drakes seems to be the most overwhelming one. the reason u have so little stats is because some of us have agreed to not even play the matchup
idea to fix: give all drake units +10 cold resis so -40% for fighter and clasher and -30 for burner

My thougts:
It's definitely a good idea and fully agree with Cremember. I would also increase cold resistance of saurians (to 0% or +10%)

Original quote:
4. dwarves seem like a harder fix basically they do well in matchups where thunders are effective(loyal and drakes) its the single best unit to get a strick
idea to fix: +15%xp cost for thunders both 1-2 and 2-3 and -15%xp cost for guards 1-2 and2-3(this can help in the orc matchup)

My thoughts:
I agree with thunderer changes, however, I don't see the point of reducing xp cost for guards.

[blizz]

here are my points

1. "I believe +10 xp cost (or something similar that gives a round number as a result) and + 1 gold increase is enough as first step". I think this is a better idea

2. I feel that there is a disadvantage for northerenrs in general. But it is not so much after the intro of leader. In my opinion, orcs vs elves really comes down to the map only. If there are far few trees in the map, orcs can win easily. I find the strategy where keeping 2-3 leaders at lvl1 and 1 leader at lvl 3(usually the intelligent one) decently effective against elves. Maybe it is good to just reduce the lvl 2->3 xp only. or give +10/+20 arcane resistance to assasin might also help

3. I like the idea of saurians with cold resistance. It will encourage players to pick more skeletons to counter

4. i dont like the idea of increasing xp cost for thunderers. It wont be effective. even 1 thunderer with 2 stikes can kill any drake unit, which you can get with an intelligent dwarf. It is especially difficult because of the melee damage which is blade. So saurians become ineffective. We need a fundamental redesign. For example, if we reduce the melee of thunderer to 4-1, 5-2, 7-2 (at different levels) then it will be difficult to lvl him up against saurians/thrasher combo and they can kill him faster. Same with loyalists, halbs/javelineer can kill them faster too (kinghts can kill them also without worrying about miss/counter charge damage)

reducing xp requirement for guard is a very good idea too. But keep in mind that it is also an underrated unit against loyalists. not easy to counter it

[Bad_tactics]

I agree with blizz above. What I would add is the following:
1. Rebels have the best variety of decent units that are easily deployed depending on the opponent. Consider taking the mage away. Also reducing the number of forest tiles on the maps would give units like clashers and grunts a better chance against them.

2. I was made to understand magical on northerner shamaness would make them too potent against dwarves. This can be countered by giving all dwarves the (boosted) healthy perk from ladder (gives chance against undead too). This way shamaness gets magical and northerner/dwarf balance is maintained.

3. Saurian soothsayer can get their ranged damage changed to arcane.

4. Thunderer is always OP. The melee NERF would be perfect. Consider doing the same to elf archer and elf shaman.