Afterlife (with XP mod) Statistics

Discussion of all aspects of multiplayer development: unit balancing, map development, server development, and so forth.

Moderator: Forum Moderators

Post Reply
dwarftough
Posts: 271
Joined: August 4th, 2019, 5:27 pm

Afterlife (with XP mod) Statistics

Post by dwarftough »

Hi everyone! In this topic I want to share stats on the Afterlife gamemode as it's played in AfterBot (rated Afterlife). This is the first stats made for any Wesnoth mode that is not 1v1.

The total sample size is 4716 valid games played on AfterBot. From them for the most of my studies I selected 1190 games where both opponents had 1600+ afterbot rating.

The settings of the game: Afterlife with Afterlife Random map picker, 100% exp modifier (default for after), Plan Unit Advancements, XP mod (4 hp = 8xp, 1 move = 10 xp, 1 damage = 16 xp, 1 strike = 56 xp, upgrades cost 1.5 times more if a unit has "drains")

A brief overview on factions

The conventional wisdom about Dwarves being op seems generally be confirmed: they have higher winrates with almost all factions, only Orcs and Undead seem to get on par with them.

The next tier of factions is Elves and Drakes, they have mostly good results except for certain matchups.

Undead and Loyalist are generally average performing factions, they are okay in most matchups but have a weak matchup as well (as we'll see in permatch stats).

Orcs are clearly inferior, they are behind in almost every matchup except undead and dwarf (surprise, surpise :D most likely because of poison).
Last edited by dwarftough on May 16th, 2022, 11:33 pm, edited 1 time in total.
Co-founder and current maintainer of IsarFoundation, Afterlife Rated and overall Wesnoth Autohost Project
dwarftough
Posts: 271
Joined: August 4th, 2019, 5:27 pm

Afterlife Statistics: Per Match-Up

Post by dwarftough »

So, general stats are not that interesting, let's see per-matchup stats, there things go interesting. We will study winrates with games of only 1600+ players.

First, the general table of winrates among factions (mirrors are prohibites in the bot, so no stats for them, but well...)
IMG_20220516_061624.jpg
Now, we can divide matchups according to these stats: Very Tough, Tough, and Fine. Let's look at all of them, starting with

Very Tough
  • Dwarf vs Elf: winrate 78,48%
  • Dwarf vs Loy: winrate 75,00%
  • Elf vs Undead: winrate 72,29%
  • Drake vs Orc: winrate 69,31%
These are very tough matchups according to stats (for losing side ofc xD). On Dwarf vs Elf, I think that problem existed even in default 1v1. Opness of Dwarf in all its glory: good resistances, while Elfs don't have any protection from blade damage which is abundant among Dwarves' faction (even Woses, they have 0% blade resist).

Dwarf vs Loy is somewhat unexcepted, I'd think Loy can fight dwarf better with chargers, mages, heal and other options this faction has. But stats say as they say. Dwarf has enough resistances to stand their grounds firmly against Loy, we have this 75% winrate for them.

Elf vs Undead, though, is not unexpected. I remember in 1v1 stats it was also like 70% winrate for Elf, although some believed it's an outlier. In afterlife some folks said that Elf is clearly superior and it's not clear what to do: mages, both with fire and arcane, as well as Woses just crushes Undead. Some other folks said Undead has powers to counter. Overall it seems playing Undead in this matchup is tricky, elvish arcane and fire mages and woses seem to be a really decent thing. So 72%.

Drake vs Orc, well. In usual 1v1 this is often believed to be favourable for Orc, especially on some specific rush-friendly maps. Orcs have cheap units, so the first night rush against Drakes with their mighty but expensive units achieves pretty good results. But Afterlife reverses this aspect. You'd better have expensive units with powerful upgrades than cheap units with mediocre upgrades. And it's the exactly this situation. Drakes units have very powerful upgrades, also their have access to healing and leadership. While Orcs' upgrades are not that great, and cheapness of recruited units have no much benefit here. After gathering these stats the balance patch for Orcs was applied, to enable Orcish Xbow to Orcish Slurbow to advance much faster (50 xp instead of 80 xp). Hopefully it may help a bit in this matchup, but we'll see after a while.
Co-founder and current maintainer of IsarFoundation, Afterlife Rated and overall Wesnoth Autohost Project
dwarftough
Posts: 271
Joined: August 4th, 2019, 5:27 pm

Re: Afterlife Statistics

Post by dwarftough »

Tough
  • Dwarf vs Drake: winrate 66,67%
  • Elf vs Orc: winrate 63,29%
  • Undead vs Drake: winrate 61,54%
and arguably
  • Dwarf vs Orc: winrate 58,21%
Dwarf v Drake seems like a tough one with Drake having powerful upgrades but Dwarf having powerful upgrades and good resists as well. As well as evasion on hills/mountains. Also extra strikes for a leveled thunderer might be painful for Drake. Overall Drake units are strong and can deal damage, but there is no clear evident plan for Drake.

Elf vs Orc, can't say much here, probably Orcs stay worse because of mediocre upgrades and lack of ranged (Elf's ranged is much more powerful).

Undead vs Drake is as controversial as in 1v1. This matchup is hard to play correctly for Drake, but not that one-sided. Although the map isn't big, which is bad for Drake, Drakes advancements are powerful. Cold damage of Undead is tremendous, but Infernos burn Undead a lot as well. So in their ToD both factions can deal devastating damage. So, it's an intense game, but easier for Undead, so 61,54%.

Dwarf vs Orc, well, this is more debatable than tough. Stats are a bit in favour for Dwarf, and some folks think that Dwarf has a clear upper hand here: his units far better than orcs. But other complain that orcish poison makes it quite hard for Dwarf to play (well, technically Dwarf can get poison as well if levels a thief up to the 3rd lvl, but no marksman and not that easy). If Dwarf doesn't handle poisioning of his units well, he can lose easily. So this matchup is more or less okayish to my taste, at least orc has a better winrate here than even against Elf.
Co-founder and current maintainer of IsarFoundation, Afterlife Rated and overall Wesnoth Autohost Project
dwarftough
Posts: 271
Joined: August 4th, 2019, 5:27 pm

Re: Afterlife Statistics

Post by dwarftough »

Fine
  • Undead vs Dwarf: winrate 57,33%
  • Loyalist vs Orc: winrate 57,14%
  • Drake vs Loyalist: winrate 56,79%
  • Loyalist vs Undead: winrate 56,04%
  • Elf vs Loyalist: winrate 53,85%
  • Undead vs Orc: winrate 51,90%
  • Elf vs Drake: winrate 50,68%
Undead is the only faction with the positive score against Dwarf. Although in usual 1v1 Undead vs Dwarf sometimes believe to be strongly in favour of Dwarf, here you can play with lots of adepts or ghosts, so Undead manages to stay good against dwarf.

Loyalist vs Orc turns out to be quite moderate, although some folks claimed Loy should be destroying orcs with chargers. Also suddenly some folks suffered from poison with Loy (although Loy has healers). Overall, stats more moderate than expected :D

Drake vs Loyalist is surprisingly in favour of Drake a bit. Most likely because Drake can deal with chargers using first-drake high-level Arbiters.

Not much can say about other matchups (really, what can you say about even matchups :D )
Last edited by dwarftough on May 17th, 2022, 1:05 am, edited 1 time in total.
Co-founder and current maintainer of IsarFoundation, Afterlife Rated and overall Wesnoth Autohost Project
dwarftough
Posts: 271
Joined: August 4th, 2019, 5:27 pm

Afterlife Statistics: P2 advantage

Post by dwarftough »

Another interesting thing is the P2 advantage. Stats show that P2 has a bit better chance of winning, for 1600+ it reaches 54%. The supposed reason is that AI makes its turn first against P1, so P1 may fall earlier. For lower ratings P2 advantage seems to be less visible, although on lower ratings all stats are more blurred anyway
Co-founder and current maintainer of IsarFoundation, Afterlife Rated and overall Wesnoth Autohost Project
User avatar
IPS
Posts: 1007
Joined: December 6th, 2009, 6:36 pm
Location: Venezuela

Re: Afterlife (with XP mod) Statistics

Post by IPS »

Well, it's not stats or statistic sharing, but general observations that explains why such win rates and why all matches even statistically.

1.- Dwarves are only somewhat more fragible to magic (fire, cold, arcane), but Dwarves have a STRONG counter for all mages or archer, and this is ULFSEKER / BERSERKER. This forces Opponent players to get dissadventageful situation of recruiting melee to counter the berserking units. Berserk also counters one of most nasty northerners units: the orcish assasin. A Single ulf and some luck can chunk an instant lv2 unit. Not counting that berserker has 4 melee strikes, which means it will benefit more from XP upgrades than average unit ...

2.- Steelclad being sincerely OP, they have insane HIGH HP and RESSISTANCES, there is nothing that can match a Steelclad or Dwarvish Lord in on at least 60% defense on hills or much worse the unstopable 70% defense ... at having TWO attack types, there is no fighter unit that can counter them (Loy Swordsmen or Loy calvary are the only exceptions) , but the problem of this ... is that Steelclad has even higher ressistances than both units (Well Loy calvary not, but has massive -20% pierce ressistances and incapable to get at least 50% terrain defense ...).

3.- The map-pool on ranked After-life in general favours the movement costs and not that much raw speed such as of Calvary. Also, you can spend very few XP on movement to make steelclad insanely faster than average unit, making a better use of 2 extra movement than almost any other unit spending 2 extra movement points. The maps are too small, and don't allow players to hide/block correctly for a too long period of 2 turns because of the insanely small maps ... meaning that in certain momment of game, opposing player will suffer of -25% combined of sick dwarf ressistances and being forced to hold positions in situations that in a normal match is a bad idea ...

4.- Combined of 3, villages and oasis in may situations are 4 up to 5 tiles of distance from the nearnest village/oasis, making poison a not a big deal at all. This reinforces point 3, topic of wrong and poor map which leads to unfair metas.

Other points and tiers are blatant ... but the main topic of this. There are MORE FAIR afterlife maps for dwarves and other defaultish factions ... problem is that the most popular but crappiest of maps is the one that is used to define player's ranked matches. There are bigger and more equilibrated after-life maps which are not used in ranked matches.

My biggest recommendation is that after a period of time, restart the season ranked with a more fair map, or even, mapping it to be more fair and less RNG dependant on what faction you get what will determinate in many situations if you will win or lose. Still, if you're skilled you will get higher rated and win matches that a less skilled player would lose, but Meta is strongly defined because of the crappy map pool in ranked match being the most notable problem of After-life ranked.
Creator of: Deathmatch new in 1.12 server.
Co-creator of: Era of Magic in 1.16 server
Developer of: Empires in 1.12 server, Ageless Era in 1.10 to 1.16 servers (but innactive recently)
Try My winning Orocia Guide
dwarftough
Posts: 271
Joined: August 4th, 2019, 5:27 pm

Re: Afterlife (with XP mod) Statistics

Post by dwarftough »

IPS wrote: August 3rd, 2022, 4:04 am There are MORE FAIR afterlife maps for dwarves and other defaultish factions ... problem is that the most popular but crappiest of maps is the one that is used to define player's ranked matches. There are bigger and more equilibrated after-life maps which are not used in ranked matches.
I would like to use bigger maps, I even tried one, but many folks didn't like big maps because games are longer on them.

But maybe the change is possible in the future, with restart or not

But actually I'm more inclined to change units than maps
Co-founder and current maintainer of IsarFoundation, Afterlife Rated and overall Wesnoth Autohost Project
User avatar
IPS
Posts: 1007
Joined: December 6th, 2009, 6:36 pm
Location: Venezuela

Re: Afterlife (with XP mod) Statistics

Post by IPS »

In case of trying another map rotation, you can make another bot that hosts a different map, I think that can be tried as a test and se how it works.

About rebalancing units for a fair afterlife match... well, quite requires much effort and force players to download an add-on era for more fair Afterlife matches. But as far, few units needs ressistances or stats adjustment, but many require so much XP for a full healing AMLA and sometimes even not worth it upgrading it to lv3 unless wanting the expected full heal (Like as of Elvish Ranger or both Lv2 drake clashers).


But most notable in far my like lots of afterlife matches, my main complaints about balance would be these:
- Steelclad batlantly OP in this map, should be -10% all physical ressitances and way less HP or more XP to lv3.
- Not worth getting lv3 of certain units like Red Mage, White Mage, Revenant because they demand an unfair amount of high XP that could be used in something more useful, even in XP mod.
- Ulf need to be 3 strikes unit instead (5-3 --> 9-3) to not make him that OP when cheesing with XP mod. More chances for rest of units to counter an over-buffed berserker.
- Undead being incapable to correctly kill Woses at time, player is forced to rush Lv3 Lich and avoid recruiting unnecesary XP that could feed enemy units againist rebels matchup.
- Revenant being worthless because Deathblade has 1 free extra strike and 1 movement, and just 8 HP less... XP too high to make Lv2 revenant into a Lv3 Death Knight , with worse ressult than farmed Deathblade XP mod. Revenant also being less weaker to impact/fire/arcane , stats that cannot be adquired by XP mod, could suit.
- Skeleton Archers being a complete garbage, they literally only exist to counter loyalist calvary lmao. Underpowered as hell, specially on Lv2 / Lv3.
- Drakes somewhat forced to spam fighters because Lv3 blade master is best option because of the "low" XP required to be obtained, unlike other units like Lv3 Clashers or Burners.
- Saurians too low in stats to be worth it, even if sometimes needed to counter stuff like Horsemen, still they have no chances againist Khanlgan Thunderers because of strong Blade melee. Agurs have Lv2 cap, whicch leads them to be XP waste if killing with them after they get lv2 because other units scales a way better with certain mounts of XP than giving extra damages to Lv2 Augurs.
- Loy calvary HP nerf doesn't help that much loy to deal with Knalga OPness, but quite irrelevant if Knalga gets fair nerfs. They work quite decently in many other matchups.
- Lv2 Loy Lancer is quite OP unit lol, but by experience it ends being more useful having 14-2 charge spanws over 12-3 because unit will rarely have a chance to 3rd strike until very late game.


There might be many other cases, but default is not the most balanced Era in Lv2 units XP management at all which in many situations, limits players options that are more poor and not that correct at all to counter certain matchup.
Creator of: Deathmatch new in 1.12 server.
Co-creator of: Era of Magic in 1.16 server
Developer of: Empires in 1.12 server, Ageless Era in 1.10 to 1.16 servers (but innactive recently)
Try My winning Orocia Guide
dwarftough
Posts: 271
Joined: August 4th, 2019, 5:27 pm

Re: Afterlife (with XP mod) Statistics

Post by dwarftough »

IPS wrote: August 4th, 2022, 6:43 am force players to download an add-on era for more fair Afterlife matches
For simple fixes not necessary, I already changed orcs, enabling the 3rd level for assassin and lowering xp threshold for orcish xbow, and it doesn't require to download the addon.

I tend to disagree with few of your complains though
IPS wrote: August 4th, 2022, 6:43 am . Agurs have Lv2 cap, whicch leads them to be XP waste if killing with them after they get lv2 because other units scales a way better with certain mounts of XP than giving extra damages to Lv2 Augurs.
Augurs lvl 2 cap means they can be led by the 3rd level leader. So I've seen people using the 3 oracles strategy, with pretty good results actually, the damage is quite-quite good, even dwarf needs to be cautious because oracles deal real damage to lords. Ofc you need several shield drakes, but overall augurs are viable. Skirmishers might be debatable, but they are skirmishers, which is quite insidious here
IPS wrote: August 4th, 2022, 6:43 am XP too high to make Lv2 revenant into a Lv3 Death Knight
A brief note, there is no Death Knight advancement enabled here
IPS wrote: August 4th, 2022, 6:43 am Drakes somewhat forced to spam fighters because Lv3 blade master is best option because of the "low" XP required to be obtained, unlike other units like Lv3 Clashers or Burners
To be honest, I can't be so sure about this, burners and clashers taken very often. One of the strats is to get the 3rd level leadership asap, and then keep you units in the 2nd level as long as possible (so you may have near advancement xp but use it with xp mod unless you think level-up became necessary). This way you benefit from leadership, provide less xp for the opponent, can level when you want to full-heal, which is a big deal, and pay less upkeep. And high xp req might be preferable here in fact, I doubt even that lowering xp req for orcish xbow was really that buff than nerf.


I agree though about Steelclad, I myself thought about -10% phys res for him and lord
Co-founder and current maintainer of IsarFoundation, Afterlife Rated and overall Wesnoth Autohost Project
Post Reply