Greedy Goblin

Tuesday, January 31, 2017

The fall of "skillz"

Yesterday I completed my promotion to gold. I'm glad that I've spent a few more days with the horrible game League of Legends, not only because "Gold" is unreachable by 75% of the players, so it's a good marketing for my don't play League of Legends page. Please note that I've not simply dragged myself to Gold with 51% winrate like the "skillz" players, but with 25 wins, 11 losses with Nunu after I've found the Thunderlord masteries. The new Warwick is banned too many times to be usable and haven't played Nunu for a month before utilizing him for the Gold rush. Here is my stat page:
Such winrate clearly shows that I am far from my plateau and much higher ranking would be reasonable goal if I'd spent more time with this damn pay-to-win crap.

Sure, it's also a plus that I've found perfect examples of all-out rigging, but the real reason why it was a good thing to do is that I can show that during my climb from Silver 2 to Gold with Nunu I was:
  • Second highest damage to enemy champions 1 time
  • Third highest damage to enemy champions 3 times
  • Fourth damage to enemy champions 8 times
  • Lowest damage to enemy champions 23 times
Ask any "skillz" player and he'll say that I'm "low" and was "carried". My teammates often told me that. Either I am the luckiest player alive, always getting boosters, or winning the games have very little to do with "skillz". Those who "carried" me had average winrates while I climbed fast. Why? Because they focused on something irrelevant: playing better, instead of what matters: beating the system.

I've sometimes used Thunderlord (not for itself but for +5% cooldown reduction) with full ability power build and sometimes Bond of Stone with full tank build. As I had to select the masteries and runes before the game started, it's clear that I knew in advance what build will work. Why? Because I know how the matchmaker rigs. If it gave me a "Sure Loss", I went with full tank and enganged the enemy early to force teamfights instead of peaceful lane game where the carefully selected enemy carrier could feed on our selected feeder. In teamfights our OKish team beaten the buyers+boosters enemy, my job was only to engage first and take the ulties of the buyers who just couldn't resist focusing the nearest red, even if he had Spirit Visage, Thornmail and Warmog. When the game was easy win, I just disappeared in the jungle and fed on monsters building a magic damage glass cannon. When teamfighting phase naturally arrived, I was 2-3 levels over the enemy jungler who wasted his time ganking, my ultimate damaged over 1500, my snowball damaged nearly 400 magic and we had 3-5 dragon advantage thanks to my busy soloing. In fair games I picked the second strategy, simply because I liked it better, sometimes worked, sometimes not.

If I played every game with Thunderlord-glass cannon, I'd lost 80% of the Bond of Stone games, keeping me below 50% winrate. If I played every game with Bond of Stone tank and early engage, I'd lost most of the Thunderlord games as good players simply ignored the unkillable but zero damage Nunu. The trick was that I knew in advance what kind of game will come since I understand how the matchmaking works.

The players with the high damage defeated the opposing players in one game. I, with my #4-#5 damage defeated the system in most games. Needless to stay that focusing my energy to increase my play would have much worse returns than analyzing the system. This is why I don't even bother to argue with "skillz" players, I simply moderate them out. Poor things are not even wrong, they simply stuck on a simple understanding of the game and unable to contribute to any meaningful discussion. Don't be like them! Understand the system and beat the system if you want to climb high. Or show everyone how it's broken and have a high traffic blog!


maxim said...

Funny how you confuse your own (very much well-earned) skill at evaluating the relative composition of teams with "beating the system", while at the same time crying that "skill is dead".

However, given your account of the kinds of players you played so far and what they consider to be "skill", i am not surprised. What i'm surprised at is how gleeful you are over your own feeling of superiority over those kinds of players. You must really enjoy what amounts to bullying babies and cripples.

As for your claims on how far your skill (defined as the sum of your understanding of the game) can take you, i'll believe it when i see it.

Gevlon said...

@Maxim: evaluating teams could be done by a script. "Skillz" would mean that I'd instinctively know what teams work on what.

Also, I'm not evaluating teams based on their merits, but by realizing that the matchmaker is rigged. "Play with glass cannon if two of your teammates are new and two are good but play aggressive tank if two are good and two are average" makes absolutely no sense. It works because I realized that "new" = BUYER and that the enemy team will be fine-tuned to lose laning phase.

Anonymous said...

I also find it interesting that you rule out the effect of learning stuff by playing and simply attribute it all to your system. Eg. we have no idea whether your warding, map-awareness and strategic decision making have actually improved..

Perhaps try to make a new account, play something like garen/mundo top, annie/katarina mid (those are all so called low skill champions), try without any team analysis. See where you get. That would create a true comparison point for you, where we could see what was the effect or your system. It would give a lot more creditability to your analysis.

As an example on how this can affect, I played only support and jungle until I got to gold. Then I thought to myself that it's only because I've learned a few champs and I suck with other roles and never played them. Then made a new account, played only mid and top....and got straight to gold. Then made another account..adc (with occasional mid)..and got straight to gold. Analysis here was that it was not necessary to have skills with champions, it is enough if one understands vision, map, items and strategic decisions.

Gevlon said...

@Anon: so you believe that map awareness and warding explains why a straight lowest damage character climbs to gold?

I do not question that a skilled player can climb to gold (actually that's more or less the definition of skilled). I'm claiming that a totally unskilled one can climb to gold just by beating the rigging.

Anonymous said...

I don't think I'm the only one not surprised by your low damage output. One single target spell with short range and an ultimate that is hard to pull off properly. Plus, you spend most of the time in the jungle, doing zero damage to enemy (while lane players are poking each other and brawling).
Not really a proof of anything.

Gevlon said...

@Anon: my champion selection and my playstyle explains my low damage of course.
But how do you explain that I win despite low damage?
Nunu is barely played and my teammates constantly whined for ganks and cursed me for being in the jungle. According to them, I did it wrong and should lose.

Andi Y said...

damage is not all. you just prove that going out if silver basic game knowledge is enough. you just need to understand the game basics about laning, teamfighting and when to do what.

you just learned how you play LoL with Macro Management. Micro Management will be required in higher elos.

i dont know why you didnt test also on another account. where you brought one skin via money.
your stats and theories are interesting

but currently it is more about your game skill. you didnt understand it yet, that you actually outskill most silvers, by knowing when to lane/teamfight/decision making etc.
this is a fact which let your elo raise. not only the champ you actually master.

High elo guys can pick any champ they wont and win in low elo. not because they master the champ better in laning phase, but they understand the macro game and have better decision makings.

Gevlon said...

@Andi: then why was I losing 80% of the "Sure Loss" games before I realized that they are Sure Loss games? Did I have no Macro Management skills?

Your idea would hold if I'd climb only by playing. But I didn't. I stuck on Silver 4 and started to move only when I assumed that the game is rigged.

Andi Y said...

maybe. i dont say you are false. the truth is in the code and we can only make assumptions.

i also face sometimes loosing streaks and winning streaks. i didnt changed my playstyle.

i dont say you are false. the truth is somewhere in the middle.

but you have to admit, that understanding the macro game, will get you to gold. New players are placed in silver anyhow. ye i hate this also, why not bronze. the first step to improve, is macro game.
and you describe very good how you improved macro wise. sure loss means you get a terribad laner in your team, so you provoke teamfights early. isnt this also understanding the game and macro gameplay??? it is.

maybe league isnt so much rigged as you make it to be. you just improved makro wise.

Darkgold said...

Nunu is designed to win games with the lowest damage by controlling objectives and the enemy jungle/ empowering his team. I'm not surprised that you did better; a lot of people play a lot of new champs or difficult champs, forcing them to learn how to use abilities every game instead of learning strategic or tactical skills. It's often suggested to play low-skill champions at first to force you to improve at actual skills, instead of having to relearn difficult abilities. It's not recommended to play hard champions until plat+.

Also, I think the matchmaking system is "rigged" to temporarily reduce the ELO of players trying new champions, so they don't get discouraged at their lowebsite champion pool, but lose whenever they pick you a new champion (because they don't know how to play) and quit by being discouraged. "Never attribute to malice what can be explained by stupidity." It's just rigged too much to make new champs win, to ensure people can always find a new champion and keep playing.

It doesn't makeven sense that new champion players would be more likely to buy- old players can reliably pay for skins, and are much less likely to quit (=more likely to pay for more skins". Business models that favor new customers over old generally run out of customers; league is ~10 years old?

Cathfaern said...

Correct me if I wrong but LoL rigging is based on that heroes can be feed. In Heroes of the Storm it can't be done because the whole teams gets XP, not only one person. I would be curious if there is some kind of rigging in HotS and if yes how they did it?

Anonymous said...

@gevlon: why you lost 80% of the sure loss games?

Because you mastered a skill that is mostly present with higher level players.

For most players the learning goes:
1. micro ("play better than the other) => get a little over 50% win chance and rise
2. meta ("pick counters/ basic map awareness / ...)
3. macro ("when to fight, make calls etc", not always present)

you went:
1. Macro
2. ?
3. ?

It shows, macro knowledge is actually very strong. In theory you could have won the games without your pre-game analysis by recognizing the patterns in game. Your method is quite unique on this out of game part.
Additionally you proved that you have a very good planning skill plus a few good ideas how to "force the shot calling".
Usually what you did would be done by communication in the higher elos and with the pros.
One person is the shotcaller and calls what teamfights to take and when. You found an interesting way to accomplish that even if your team is not listening.

As such: yes you didn't have macro management skills before, since you could have seen the right decision in-game before.

That beeing said: your observations are a great way to show people why macro is amazingly strong in LoL. And that is not a widely known fact.

Gabe said...

I'd love to see how this works on Heroes of the Storm.
As much as I see many elements are similar (sometimes I lose/win unexpected), but some of them are not or not exist (roles, items).
Still curious, if there are some similar method to change the game?

Gevlon said...

@Andi: then forget my personal improvement and on the exploit mode and focus on the really important stuff: that new champion users (buyers) get boosters and a tailor-made team that is meant to lose

@Darkgold: the matchmaker places the PLAYER into the game, then the player select the champion. The matchmaker can't know that the player will pick a new champion - except if it uses the info that the player recently PURCHASED the new champion. Unless you know any other mean for the matchmaker to predict new champion usage, your claim can be understood as "matchmaking system is rigged to temporarily reduce the ELO of players with recent purchases". This is the most pure definition of "pay to win" I've seen.

@Cathfaern: I don't know, nor I really care getting into a "HotS is rigged" project.

@Anon: And I mastered this skill overnight? I mean I struggled for over 500 games and then bang, I jumped 4 leagues with clear expectation of further climb.

Reacting in-game is too late. I have to pick proper masteries, early buys and act before the problem happens. Even worse: if I choose to early engage, I can ruin an otherwise won game! This is exactly why the matchmaking is rigged: I'm not reacting to the game but use out-of-game info (the data) to predict something that it shouldn't predict. I mean consider two games, both game has a Yasuo top, with the same player playing it. Bottom doesn't roam top, so whatever they do doesn't matter to the performance of our Yasuo. Still, if Bot and Supp are new champion users, our Yasuo will go 10/0, while if they have 10+ games with near-50 winrate, our Yasuo will go 0/10. Explain this without rigging!

Anonymous said...

Psych 101, if winning does mean that your project failed, you sure as hell are going to lose (more often), if you think you found your grail this barrier vanishes. Alternativly believing you can do it... 1000 movies told us already.

Psych importance should not be underestimated (see tilting).

Anonymous said...

As several commenters have pointed out. Your game skill has significantly improved, at start you dont influence at all the game, now you use before you even start the game. Because you know how to use uneven lane matchup as on your advantage, your winning chance rises.

Im still convienced that matchmaking is rigged, or even if its not. Relatively random teammates, happens often in plat too. Everything you done perfectly explainable in non rigged matchmaking system, just take into account waiting time, elo rating and ranks. If you put everything together, it can prove that matchmaking is doing bad job at best. Fixing uneven matchups with algorithms is harder problem then you could see here. Literally, they have been changing champion select systems many times in recent seasons just because of those same problems you face.

maxim said...

"Play with glass cannon if two of your teammates are new and two are good but play aggressive tank if two are good and two are average" indeed makes absolutely no sense.

On another hand,"In the event of your team having two weaker players against a team of superior players, you want to shorten the laning phase as much as possible so that the enemy doesn't gain too much of an advantage" makes perfect sense and doesn't have to involve "rigging".

Here is a metaphor for you. Francis Bacon and company created modern science because they irrationally believed in God and thought this would be a great way to get close to understanding the divine. However, science itself has very little to do with God and works best when practiced in a secular way.

By the same token, your pursuit of rigging for out-of-game reasons did indeed lead you to explanation of this particular set of play tactics. However, the tactics themselves do not require rigging to be explained.

maxim said...

You didn't master a skill overnight. In fact, it took all of those 500 games for all elements to fall into place. However, it is not uncommon for a certain skill to be useless until all of its elements are into place and then suddenly become useful once everything clicks together.

It is like a medical syndrome of sorts. Things don't get really bad (or good) until you get the perfect combo.

Gevlon said...

@Maxim: Amazing! You rather believe that I'm an awesome player (despite my personal stats are crap) than accept that the matchmaking is rigged.

If I see 2 total newbies, every reasonable person would assume that they will lose their lanes badly and the best course of action is to shorten laning phase. Which is the opposite of truth: I should make laning phase longer because the team has carriers planted who will massacre the opponents. If I see average history players, the reasonable assumption is that they will perform in an average manner. Instead they are massacred unless laning phase is shortened.

What you have to explain otherwise than rigging is "how come that new champion users aren't smashed but smash their enemies"?

Skeddar said...

I'm not quite sure why, but it seems people don't get your way of argumenting. Maybe it's best to set personal accomplishments aside (cause you COULD have gained skill overnight or the new patch favors Thunderlord Nunu players or whatever) and just focus on the matchmaking itself.

Every LoL player has experienced 20 minutes games which were very one-sided. That can either be explained by pure chance (one player made some risky moves and miscalculated, leading to a fed carry for the opponents) or poor matchmaking (bad programmed or rigged). Now your observations that people who have no known history with certain champs but don't feed is interesting... the only 'proof' missing is to compare people who owned a champ for a long time but never played him and people who invested money and got a new champ/skin. Which is probably impossible on a large scale.

Actually the most 'alarming' observation, which I also can confirm, is that dodging games leads to gigantic loss streaks, although your teammates seem fine in champ select. I don't think it's bad punishing dodgers (like with a time ban or -LP), but rigging their future matches WITHOUT ANY OFFICIAL NOTIFICATION just seems unfair to the player. Sure, trying to cheat the system isn't great, but in any other case you get a notification with a known punishment (low priority queue, chat restriction, account ban). If they rig matches without telling anyone because you broke some rules, I don't think it's that far fetched to assume they rig matches for various other reasons that benefit their company.

In summary, maybe the best way to convince people is to analyze profiles and ingame performance (why did Alice perform well with a first-time champion against Bob who played his main with a great win rate) instead of your own performance, as people either call it skill, luck, or influenced by other factors (your psychology/attitude, your decision making in the lobby, you played bad on purpose in the past and rigged your results...).

Also the whole 'Riot doesn't know which champ you will pick while matching you' seems just to add to the confusion. If Bob is a Darius top main and just bought Yasuo, what kind of lane opponent will he get? Someone who is three divisions lower? What kind of teammates will he get, will the opponents have toxic players or afklers? So I feel analyzing both teams after the game and proofing every second match was decided in champ select can lead to either 'the matchmaking is bad' (as a good matchmaking should provide both teams with a roughly equal winning chance) or 'the matchmaking is rigged'. And although creating a matchmaking is a challenging task, I think with enough data the second one will appear more likely.

Gevlon said...

@Skeddar: I'm between a rock and a hard place now. On the one hand you are right, pure and statistics untainted by me any way should be the best proof. Hey, the fact of win-after-loss and loss-after-win is more likely than statistically should be is enough proof of rigging. The win-with-newbies should be red flag.

On the other hand 99% of people are shit at maths and cry "lol tldr numbers XD". For them I must play and win with the exploit. By doing so, I am present in the results and they can claim that I merely won by "having skillz". I failed to find real-time statistics and past-time statistics are useless as I can't tell if the player was new champion user back then or not.

If someone just bought Yasuo, he'll get someone with bad winrate as top opponent who will likely be clumsy. The formal leagues will be equal, but remember that it's PLAYER and not champion league. The opponent can be the best ADC in the world if he is autofilled top. New Yasuo will get good teammates whose main counter the opposing lane players main so they can feed and carry his dumb ass. Or... the matchmaking just get full retard and give an unranked opposing player who doesn't know what lasthitting is, the enemy team also contains someone with 1000 ping and a toxic guy spewing insults on chat while feeding.

Anonymous said...

Gevlon...You talk an awful lot about how much the game is stacked in favour of recent buyers...

But you *never tested it*.

Even going by your own formulas (which may be flawed) - you had recent buyers on some of your sure-loss teams - there were just *more* of them on the other teams. Possibly by coincidence, possibly not - but we'll never know because nobody bothered to actually perform the research/tests to find out.

And it isn't like it would be hard for you to purchase a new hero once in a while to actually *test* your claims - you have spent money on games before. But you seem to prefer empty accusations with no attempt whatsoever to prove them.

I am disappointed.

Skeddar said...

"Hey, the fact of win-after-loss and loss-after-win is more likely than statistically should be is enough proof of rigging"

Hm, not quite sure about that, as in theory if you lose a match you should get worse opponents as your mmr drops, which makes it more likely for you to win and the other way around.

Maybe just as a motivation, here Riot states how pure skill and no money decide the matches:
(value #2 and #3)

Gevlon said...

@Anon: That test would not only result in giving money to scammers, but also fully anecdotal. Did I just win 3 games in a row because I payed or because I'm good. Sure if I pure $10-20 before every game and win 100 times in a row, that would be something. I'm not sure I want to give them $1000-2000.

@Skeddar: the shameless little liars. They have the nerve pitching as completely fair, when they are probably the most literal pay to win I've ever seen.

Anonymous said...

Since when you need real money exclusively to buy champions in LoL ? Nowadays, it is even easier to get a champion with the hextech crafting.
By now, you must have thousands of IPs (ingame cash) and can afford dozens of champions in 450/1350 IP tier.

maxim said...

You have demonstrated a superior grasp of strategic-level mechanics, compared to pre-gold level people. Not sure why the weird tone, it is something you definitely achieved. However, people achieve stuff on that level all the time, it is not something amazing.

I would consider you amazing if you refined that grasp into something worthy of top tier play and also trained yourself in basic twitch mechanics. However, you have, once again, chosen to consider yourself above the challenge while rationalising it by some "rigging" excuse.

Judging by your most recent post, you have already realised at least some of the folly of this way and arrived at the correct conclusion - play the actual game, engage the actual people and don't worry about meta stuff too much.

Gevlon said...

- when your teammates are winning their lanes, let this happen
- when your teammates are losing their lanes, stop this from happening
This is hardly superior grasp of strategic level mechanics, even for pre-gold people. My find isn't that. My find is that "if your teammates have no experience with their champion, your OTHER teammates will win their lanes", which makes absolutely no sense unless you assume rigging.

maxim said...

At this point, i feel we are arguing semantics. So let me just clarify the points here.

When i say "superior", i mean it in the sense that even the most unathletic healthy human is a superior athlete to a baby. It has been my experience that in ladders the "gold" level is the highest level at which you still meet those who are essentially babies in terms of their ability to grasp the game.

The situation of "if your teammates have no experience with their champion, your OTHER teammates will win their lanes" makes some sense to me, because of how most team matchmaking designs work. Specifically, the team matchmaking algorithm will often be under all sorts of pressures to put together a team of a specific average level, thus often ending up having to supplement weaker players with stronger ones in order to satisfy the team's expected average strength.

Your argument is essentially that these pressures are created by the intent to rig the game in favour of buyers. I saw many examples where the simple time pressure of having to give everyone a certain quality of matchmaking service is already in itself enough to create such a situation.

Gevlon said...

@maxim: no matchmaking is perfect. This means that there will always be games strongly stacked against Joe Somebody and games strongly stacked for him. That's a price we must pay for shorter queues. We see it in WoW battlegrounds where essentially the number of healers in team decides the game. However I never saw it being biased. Sometimes I got lucky, sometimes unlucky. Over lots of games the odds evened out and only my skill mattered.

In LoL the games are always stacked for the buyers, therefore against non-buyers. Textbook pay-to-win, this time literally pay to win a match.