Page 1 of 1

League Theory: Format

PostPosted: Fri May 11, 2018 3:43 pm
by Glas Masv
|||This post is the product of ruminating on factors that fundamentally affect player performance in league and I have drawn novel conclusions. These as well as other original insights are stored together in my guild's steam group forum, Chaos Reborn: Hegemony Pantheon. Readers who wish to learn more about guilds or how to join I direct to the sub-forum, Orders and Guilds. In writing this a thread a normal post became an essay and ended as a lecture in length and tone. I think this writing has too much heuristic value to keep confidential or hide in Orders & Guilds. I'm not claiming that everything I'm opining is true, but I think it is worth reading and more so as much as the reader has experience playing Chaos Reborn. My aim is to introduce you to factors of the game, league play specifically, that you often interact with, yet are rarely discussed if ever conciously acknowledged. Intentionally considering these factors has the potential to radically affect your performance in league, but I'll settle for simply causing you to think about the game in a new light even if you disagree with my conclusions. Enjoy.|||

League Theory: Format

I've been cogitating on how the process of selecting formats and the size of the set of league games played in a month both affect Rank Progression and Rating. I don't know how the rest of you queue for games, but I imagine you select mostly 3 and 4 player games in your chosen mode & rules. I imagine you also queue for a sizeable stack of games to bring your total up to a chosen threshold and do this at the beginning of the month when new seasons begin. Some of you may top off your queue over the course of the month replacing finished games by starting new ones or see the active game count dwindling and re-queue for a stack of games in the latter weeks of the month. Others may stick to their original game count for the month and if its not going well lay off league to wait until a new season. This 'Format' theory will explore the different dynamics in approaching format, game count selection, and re-queuing.

I'll start with game count. I have seen players win gold in Equipped Chaos having played just 14 games I believe. Egeon did it and won with a 1500+ rating one month in 2018 I if I recall correctly. He must have signed up for mostly if not all four-player games and benefited from a strictly limited game count, or more likely only 14 or so 4-player queued games filled up with 4 players. Further, he won often enough to hit Archmage which means he was red hot that month. That is an atypical performance in league. Since 4 player game queues don't fill up quickly there is a limited number of them active at any one time or total inside a month. If Egeon had simply done well he likely wouldn't have hit Archmage in so few games; only red hot performance could manage to make Archmage. I'll get the 'why' in the next paragraph when I discuss format. I'll get to the 'why he stopped at around 14 games' in the subsequent paragraph. Most players seem to favor 3 player and a mix of 2 and 4 player games. Assuming Egeon's performance is atypical it is correct to understand that there is a minimum number of games you must finish in a month. If you play a mix of 2/3/4 multi-player that number could be around 25 or 30 or even 40+ games per format. The more games you play the greater the capacity to generate rank progression. The exact minimum game count threshold will depend on the number of players and how often you realistically expect to be able to win. The more games you sign up for the greater your capacity to cushion losses and still hit archmage.*

Now I'll describe the dynamics of 2/3/4 multi-player format in order. Two-player format has the most screwy dynamics. You only get one star for risking one star. You only start losing stars when you reach Rank 5 known as Manticore Pimp or whatever its known as. Since two-player is on average much faster than the other formats and you can't lose stars a viable strategy is to stack 2-player for week 1 & 2 and diversify once you reach rank 5. The idea is you'd rather not waste time to get stars that are gauranteed by simply playing enough and winning at all. After rank 5 two-player is not recommended unless you have a nasty build specifically designed for two-player and you expect to win frequently. If you are winning 50% of the time you are not advancing. If you lose twice in a row then normally you'll need 3 or more games to simply return to your original star count. If you are like Yogicfire and spank two-player with ease then two-player turns out to be the most advantageous format with which to advance in rank progression. The hotter you are the fewer two-player games you'll need because you'll need less rank progression capacity. The key is Win Streaks! If you are hot you can play one two-player game in a smaller number of turns and win three stars against a single opponent. If you are simply doing well you'll need more games and a losing streak can really set you back because its harder to recover stars in two-player format compared to the others. So if you win 4 or 5 games quickly in a row you'll net 9 or 12 stars in a matter of days. You can burn your way to archmage in days if you are hot in two-player. Here is the big BUT.. you effectively have to start your month over once you hit Archmage because two-player victories and even win streaks are poorly rated. You'll also have less time to start 3 & 4 player games and still finish them within the same month and what's worse any excess two-player games that are still active can only waste your time or detract from your rating. The only way to avoid this is to have 1500+ rating, cut queue, and coast to gold - but its hard to get rating that high with two-player wins. Simply hoping no one else hits Archmage that month isn't going to cut it most months. This paragraph was a long way of saying for that most players I do not recommend qeueing for 2-player games at all aside from using this format to grind rank 5 if you can stomach the endeavor. Two-player is a more reliable option in Law Mode.**

Three-player format is bread and butter for most league players. Three-player is a balance of most dynamics. You can play them faster than four-player and win them more often. While in rank-progression you don't have to have win streaks to advance. You receive two stars to win for the price of a two-player loss. That means 50% is minimum to advance. Win streak bonus is still one extra star as I believe the maximum number of stars you can win in one game is three. For most players three-player is where league is at. You have a somewhat reduced win rate, but the star rewards mean that you hit even at 33%. Get over 33% win-rate and with a high enough game count you are guaranteed to eventually hit Archmage. It doesn't matter if you hit Archmage on the last day insofar as winning a medal one would think. Hitting Archmage earlier is usually better because knowing your rating and how it compares to your opponents is helpful in choosing which formats and how many games you should be playing to have the best chance of making Gold. However, if it is late in the month and say you got a late start so you are just hitting Archmage on the last day.. and say you were hot that month, then when your rating posts it may be the highest among the Archmages and your opponents, who rested on their laurels or ignored you in gameplay as they assassinated each other, don't have time to focus on you or create & resolve wins to surpass you. The problem with knowing your rating is that everyone else knows too. Not always the best to post a rating first and early in the month - food for thought. The three-player victories in rank progression remain useful as an Archmage and are rated decently. The rating does not care if you win consecutively. The rating does not care much that you lost a billion games. The rating cares that you won like 39 times out of a billion and they were 3-player games. The higher your rating the more it goes down if you lose. It seems that a two-player loss is as a four-player loss as far as the rating is concerned. The dynamic this fosters is that if you are in the lead and insofar as you are in the lead you have more to lose by playing games and less to gain by winning. It's very hard to get 1500+ rating, but if you do and you don't tailspin you can fight to get high 1400+ by the end of the month for a win. If your rating differential is comfortably high then cut queue while knowing wins and draws are desirable and committing suicide to kill your ratings competitors in second or third place - is desirable. If you are losing, insofar as you are losing, you'll need to play more games to create the rating-creation-capacity*** with which to overtake the ratings leader. Another way of saying the more games you play the more cushion you have for losing and still advancing your score enough to surpass competition as long as you recall they can work against you when fending off competition to retain a lead. If you hardly ever win four-player then stick to three-player games at Archmage as your main format. If you can win a four-player more than rarely then when you're in reach of Archmage and no further than week 3 at the latest then I recommend stacking four-player in queue.

Four-player seperates Gold medal from "Dammit, I was so close". It's also a false friend in rank progression. You risk one star and receive no star bonus for win streaks; though it counts towards win streaks with a subsequent three-player game. You risk one star for three and your chance of winning four player is frustratingly low if you're like most players. You probably have a much higher chance of winning three stars on your consecutive three-player victory and your four player losses interrupt three-player victory win-streaks. The three player games are often quicker. (Though a subset of four-player go quicker when two killers dispatch their counterparts in an early face-off turning the match into a two-player sans Magic Bolt. If you are confidant in your three-player game and think you can do better than the 33% win-rate for advancement necessary for rank progression in three-player then I recommend mixing in four-player during rank progression. I queue 3 three-player and 2 four-player games per build and queue only one 5-stack each day during Week 1. This isn't enough staggering to mean I won't get a proper cycle, but limiting daily queue stacks to 5 means I'm not taking as much advantage of gaining first turn for my games via stacking. Insofar as you stack games in queue you are giving opponents less of a chance to start first. It's an unfortunate dynamic so try not to take advantage of it more than you need to create rank progression capacity sufficient to hit archmage or rating creation capacity to win gold. Queuing for 3 three-player games per day isn't breaking the game. Queuing for 2 four-player games per day and then focussing more on four-player once you hit Archmage is a good strategy and I recommend sleeping at night even though it unfortunately has potential to break the game. If your opponents are being savvy and queuing for more four-player games than are active also then there is no problem. Besides, first turn in four-player is the least decisive format with which to have first turn. You play first in four-player and have to hope 3 other players don't disbelieve or magic attack or cast something way better that makes your heart sink into your guts. The benefit of queuing and winning even 1 in 4/5 four player games during rank progression is that if you can hit Archmage then you can count on a significantly higher rating than if you stuck to three-player format for every game. Whether you queue for four-player during rank progression depends on the difference between 'frustratingly low win-rate' and 'infrequent win-rate'. I said in the beginning its a false friend in rank progression, but it's your best friend as soon as you dispense with proletariat wizard-tier rank progression. If you are like Egeon, a psycho serial-killer for a month, then you win as soon as you hit arch-mage relying on the four-player format. If you are like normal sane civilized wizards or like Egeon is most months then four-player slows you down in rank progression. If you end a season of league just a star or two shy of Archmage then you might stop blaming your luck and point your finger at your decision to queue for four-player games during rank progression. These dynamics hold true no matter the mode. Even in Law Mode you are taking risks in the meta-game.

|||Lecture is completed in the next post due to a 20,000 character limit per post. About 3,000 characters to go!|||

League Theory: Format -- Part 2 of 2 --

PostPosted: Fri May 11, 2018 3:44 pm
by Glas Masv
If you are like me and queue for a stack of games at the beginning of the month then there is a cycle. All stacks cycle. What I mean is that some games end quickly and some take awhile. What happens depends on your build. The larger the stack the greater the impact and potential for win-streak sweetspots or interference in the early game depending on the inherent tempo of your build. If you have an aggro build then shorter games are supposed to be victories more often than losing and you'll have more shorter games. If you have a long-game build that needs time to develop to hit win conditions like megas or activations then there WILL be a set of early game resolutions and they are more often losses. Those early game losses come in one after the other in the first week or so. They retard your rank progression if you are above rank 5. They aren't as bad if you are starting out at rank 5 and just count as voided games. Losses in any format aren't a big deal, but will hurt your rating which only matters if you ever reach Archmage. Aggro builds starting at rank 9 speed through the guaranteed advancement above rank 5, but then hit a wall as early game resolution drops off and they wait for longer-game resolutions and contend with long-game builds. Aggro builds starting at rank 5 as a reward for hitting Archmage the previous month benefit mightily from the subset stack of early game victories. That early-game subset also has a higher probability of win-streak outcomes and all those stars should bounce successful aggro builds into the lower ranks. Aggro builds starting from rank 9 may want to regularly re-stack the queue of games to keep riding the cycle of early game victories, but will be limited by the drip of newly active games that seems to get slower as the month advances. Further those long-games will create losses that hamper earlier gains so after an early-game bounce Aggro builds that aren't generally dominant are racing the clock to win before long-game losses keep Archmage out of reach. Long-game builds are theoretically superior to wizard-centric builds in three-player and especially four-player league. Aggro is famous for one-time win conditions/cc and a limited number of win-conditions. If they exhaust their options on their first kill they have few safe ways of killing the other wizard(s) and as turns pass the long-game builds bloom and their safe win-conditions multiply. In practice, Chaos league meta swings from mounted-Aggro, commando builds to transmutation hordes, pricey megas, etc. and back again to counter what was scary the month before. If you use one build and you win often, especially in the first two weeks and particularly if you are Aggro - then expect adaptation on the part of your opponents. If you stacked a ton of games with your one killer build then your opponents may have had a much smaller stack and could re-stack their queue with adaptive builds to counter your threats. I'll now point out that you need to plan your re-queue stacks carefully during rank progression. If you re-stack right after the early-game sourspot is over in your long-game build then you've just scheduled another sourspot that will come right as your longer-game winning sweet-spot is in progress. The point of these sweetspots and larger stacks is to produce win-streaks during rank progression so re-queuing ought to be post-poned until the long-game sweetspot is underway. It's a bit trickier to schedule your re-qeueuing stack for Aggro builds. You should rely on a large stack initially so as to crush your early-game sweetspot and come close as possible to Archmage. If Aggro is lucky some of the slower games goes its way. If not, it needs then Aggro needs to be patient and eat its losses until fewer remain. Then re-stack and capitilize on the re-stack cycle to hit Archmage. Once Aggro promotes itself out of the doldrums of rank progression it is at somewhat of a theoretical disadvantage to improving its rating. It is less favored in the cream of the formats, four-player. Two player is worse than worthless. Often, to be safe, Aggro avoids four-player in rank progression so its further behind rival Archmages victors of four-player games during their own rank progression. My opinion is that players should alternate from Aggro to a long-game build as Archmages. At least switch to long-game builds for the many four-player games the former Aggro player should suddenly stack when becoming an Archmage. The theoretical does not apply necessarily to players who are reliably dominant in league with their Aggro builds though I contend that if they fashioned a proper long-game build for four-player stacks they might secure Gold rather than Silver during the final days when the top ratings are tight. Also, to avoid confusion - all builds that are not Aggro are long-game builds which only vary in power as it takes turns on average for their power potential to mature. Aggro builds are designed to eliminate wizards by turn 5. Long-game builds may develop by turn 2 or before turn 5 when sprites are available or with luck and so on, but they normally take longer than turn 3 or 4 to mature in power potential and none of them are designed to reliably eliminate their first opponent by turn 5. Mounted-wizard centric builds that are not Aggro exchange early game lethality for the durability to survive close encounters with multiple wizards. They lose their advantages if creature-centric builds are allowed to fully mature and, like Aggro, are more easily focussed by multiple opponents. The mounted-wizard centric builds thrive more than any other build on opponents savaging each other, stunting each other's development and wasting cc counters, etc. It's often true I think that if long-game creature-centric builds when squaring off on turn 1 simply ignored each other and minded their own development the first 7-8 turns before conducting war then no other kind of build could beat them. I'll end this paragraph on queue dynamics by observing that builds that are a jack of all turns when winning lack sweetspots for win-streaks and require more effort during rank progression. The upshot is that it hardly matters when a stack is re-queued; all that matters is that enough games are queued to hit Archmage and improve rating. There are hardly any builds like this though. Since almost all builds have a cycle when they stack games into the queue then you may as well aim for builds and relevant stack sizes that create fat reliable sweetspots to sail through rank progression. In Archmage switch gears and try to secure the re-stack queues to resolve sweetspots inside the month. If Aggro, then if you are cunning you can schedule a sweetspot before the end of the month and eat the sourspot of longer-game losses during the beginning of the next season when losing is immaterial. If long-game, balance patience with logging in more often and secure your sweetspots within the month as well as all your games within the month. It's annoying to have long-game resolutions create wins at the beginning of the next month when there is no sweetspot and therefore less potential for the win-streaks desired in rank progression. This observation is more of a guideline in scheduling and your scheduling must respond to the particular context of your ability to secure a Medal thereby foreshadowing to all our one, inexorable, ultimate future: Total Control. For Hegemony!

* (Of course, you need to be honest with yourself about how much time you are willing to commit to playing. Whatever time zone you are in I recommend logging in and tending to asynchs everyday in the evening. That allows you to take the following day off completely with ample time to fit asynchs in the second day. If you just log in during the morning that will generally mean logging in by the following evening or else you risk timing-out and losing many games if you oversleep, forget, busy, etc. Your frequency of logging in affects speed of gameplay such that more log-ins leads to more time for queuing games that can finish inside a month and thus greater capacity to hit Archmage.)****

** (Whereas Chaos is the win-rate mode, Law is the consecutive win mode ala win-streak mode when playing league. The entirely different standard of a build's worth in contradistinction to a player's skill is the primary difference between Law mode and Chaos mode; not the presence or absence of chance or even the addition of a health bar nor the differing values of spells/units. This topic deserves its own polemic.)

*** This chimerically hypenated beast sounds like the english translation of a german word. Anyone know what german word has meaning resembling this?

**** (If you are evil you might consider slowing your log-ins to a full two days per turn when facing Archmage competitors so that games don't resolve inside the month. This denies unfavorable rating changes and would make a big difference in 4-player games that would otherwise resolve to your disadvantage. I mention this for several reasons.
#1 I am evil. #2 So you can look for other ratings competitors trying to pull this on you and shame them. #3 As a warning. One of the only two requirements to be in Hegemony is sportsmanship. Manipulating ACTIVE games by delaying turn completion so Active games resolve after the season ends in order to gain an advantage is wrong. If you do it; leave the guild first. :-)

Re: League Theory: Format

PostPosted: Fri May 11, 2018 4:09 pm
by Glas Masv
It occurs to me that while most of the lecture pertains to Equipped rules it will apply to an extent to Classic Chaos. In Classic builds are by and large long-game inherently. A weak player who frequently dies early or a stack that is large enough both should lead to sourspots in Week 1. Most players and those stack sizes that aren't huge won't have enough early losses at the same time and at small stack sizes or all but high-performance players - the wins & losses will spread out enough over the month such that sweetspots aren't as likely as Equipped. Player strategy in Classic may increase probabilities of sourspots and/or sweetspots depending on effectiveness and reliably winning in the same interval of 5 turns OR ineffectiveness and reliably losing around the same time. Large enough stacks are more likely to cycle and cycles are more useful for rank progression. Ok, I'm done. Logging off.

Re: League Theory: Format

PostPosted: Fri May 11, 2018 8:11 pm
by Yogicfire
Some very interesting points. I would just make a few suggestions:

1. You talk a lot about stars but not so much about ELO. What is the difference between winning or losing a 3-way game and winning or losing a 4-way game? You seem to be suggesting that 3-way is the best choice because "it doesn't matter how many times you lose" but you are obviously going to lose more in the larger games and that will affect your ELO and stars.

Solely from an odds perspective, newer players may be better off going for 1v1s and getting a winning streak going to make the medals. If you have a 25% chance of winning a 4-way game then how are you ever going to get a run going in order to get a medal? (Plus this is probably not even 25% if you are up against experienced players.)

2. Egeon didn't play 11 games he had 11 wins and 11 defeats as I recall. So, he played 22 games not 11.

3. You also miss out on the carry over. Players don't just start a month afresh and anew. Players generally start games and carry them over to the next month. A good example of this is Ish in Equipped Law this month. After just one week he made the table with 11 wins and no losses! It will be interesting to see how many more games he plays from now on..

4. Overall, my advice is to mix it all up. Just playing one format will generally not take you to the top of the mountain. You need to play them all and WIN as many as you can.

Re: League Theory: Format

PostPosted: Fri May 11, 2018 9:02 pm
by Glas Masv
Ah, thx for the correction. 11 and 11 allowed Egeon to surpass your ELO I called rating even though you had way more victories, like over 30. The impact of losses changes depending on how high your rating is. The best way to get the highest rating are four-player victories and this is true even if victory is infrequent. I do recommend mixing multiplayer formats up, but a player that plays more four-player can win more four-player. You have to win very often focussing on just three-player after hitting Archmage in order to surpass a player logging infrequent sustained four-player wins that season. Two player isn't normally sufficient to win gold if there are multiple Archmages in contention.

There is no one way to select formats, game count, or re-queue timing during Rank Progression or Archmage. The best course depends on your build, tempo, and ability to win four player more than rarely.

If you start 3-player or 4-player with mere 2-3 days left in the month then your effectively starting the next season before it officially starts. You should keep in mind that these carryover games are played for stars if you win, but might hurt your rating in the present month if you lose very early. If you time your re-queues for week 2 or 3 you should have time to finish two stacks of games at least. Carry-over games that are slow to resolve either win you stars or have no effect assuming you are rank 6-9. Some players may not queue multiple games at once and just maintain their game count by starting new games whenever required. That will decrease sweetspot potential which affects rank progression. It will also lead to more carry-over games and format selection may be less intentional and productive.

Re: League Theory: Format

PostPosted: Fri May 11, 2018 10:34 pm
by Yogicfire
You can always select or reselect formats and games.. Everything is up in the air. In some of my games I have been waiting 16 days to start. This is not uncommon given the amount of players in league. I can just delete some games and change my quota as I like depending on need.

Also, 1v1 games go a lot quicker so that has a bearing on everything. You can get your stars by winning five 1v1 games instead of waiting forever to win your 4-way game.. Mix and match!

A lot of people don't even make it to the medals, Glass! They don't get on the table!

Re: League Theory: Format

PostPosted: Sat May 12, 2018 11:33 am
by Yogicfire
PS Most of what you say is really for Chaos league not Law league.. In Law, there are not so many players and hardly any large games are played. Ish is now on 17-0 in Equipped Law and I am fairly sure that he has just been fighting 1v1s.

Re: League Theory: Format

PostPosted: Mon May 21, 2018 12:20 pm
by Glas Masv
Everyone that isn't rocking 17-0 would do better to abandon duels in Law mode because the imbalances favor certain kinds of builds and those builds might be better countered in 3 or 4 player. Otherwise, consider copying Ish's build for duels. You can figure out how it works and what gives it problems. Then create a counter-build.