question-circle Emerald Legionnaire — First Black Chantry Nerf?

04 Apr 2019 09:21 - 04 Apr 2019 09:24 #94370 by Bloodartist


:lol: :lol: :lol: :lol:

'Lies, damned lies, and statistics'... But really, I'm just an amateur with delusions of knowing how to work an Excel file :)

I would rather present the full data so everyone can make their own analysis and conclusions, but this forum is far from ideal for that (do you know how hard it is to post a large matrix here? You have to hand craft the BBCode, it's almost like going back to the 90's... if anyone has any suggestions for an easy way to convert Excel tables to bbcode, I'd be happy to hear them. Then again, they get all mangled up by the fixed page width anyway...).

So, I will post a link to the full updated TWDA report as soon as I can.


The point I have been making (repeatedly) is that regardless of the methods you use (which may or may not be valid, its irrelevant), using them to analyze TWDA results in entirely useless, and worse: misleading data. In order to get meaningful data you must use data from ENTIRE tournaments.

Hypothetical situation A:
Tournament has 5 EL decks. Only one of them has positive winrate, but proceeds to the final and wins it. Looking only at winner makes it look like a big winner.
Reality: on average this reflects poorly on EL decks winrate.

Hypothetical situation B:
Tournament is not won by an EL deck. Oh, this must lower the winrate of an EL deck in the grand scheme of things according to you.
Reality: no EL deck participated in the tournament = the tournament doesn't affect EL decks winrate in any way.

And so on and so forth. All of this is information that study on exclusively TWDA pointedly IGNORES.

If we had data from entire tournaments, which would show that an EL deck almost always proceeds to the final (amount of EL decks participating vs amount of EL decks in the final) then that WOULD be sufficient grounds to discuss nerf/ban on EL. But that is not the data that is being gathered here.

Knowing that most people don't understand mathematics, I know how easy it is to mislead people by presenting mathematical formulas and big words. This is what is happening here and I am vehemently against it. This is why I fucking HATE statistical analysis. Its only as good as the person interpreting it, and its unbelievably easy to fall into the pit of believing false results are true, because you are using valid methods on invalid premises. The main achievement of statistical analysis is misleading politicians into making decisions they wouldn't have made if they actually knew what they were doing (see brexit).

I don't want black chantry to make any decisions on nerfs/bans/errata on false information. Start collecting ACTUALLY VALID data and then continue with this analysis. Start collecting decklists from all tournament participants so this analysis actually has some value.

ps. I know you don't know what you are doing, because you wouldn't have gone this far had you realized the data is erroneous. If you want to practice your toolbox and collect course credit or whatever, you can do so, just try not to mislead readers of this site while doing it.

A heretic is a man who sees with his own eyes.
—Gotthold Ephraim Lessing



Last edit: 04 Apr 2019 09:24 by Bloodartist.
The following user(s) said Thank You: Beatrice, DJHedgehog

Please Log in or Create an account to join the conversation.

More
04 Apr 2019 12:31 #94371 by DJHedgehog
Does the Archon allow for recording decks?

It was mandatory to provide deck lists in Netrunner events. Then again, the only thing that was recorded was the Corp/Runner for each player. Vtes has a lot more subjective decisions when you talk about a deck and it's goals in game.

For example, I got in an argument with a guy on facebook where he said a Mata Hari deck was a ravnos deck. The deck only contained 5 ravnos in the crypt (all mata hari). Would you say that's a ravnos deck? I wouldn't, and I don't know that everyone would agree. That's one of a million possibilities which can include the type of deck (Stealth bleed/wall/combat/etc).

I think, as far as good data goes, this game is kind of borked. Seeing decks that win is interesting but (as we've seen in american league) seed in the finals is a huge factor too that is not being taken into consideration. If we can say at a normal table everyone has a 20% chance of winning, the finals seeding has to skew that number significantly.

Lots of factors and as Bloodartist stated, the data is far from complete and it gives only a key-hole view to what is being played at tournaments without any context so it would be far from statistically relevant.

Please Log in or Create an account to join the conversation.

More
04 Apr 2019 12:57 #94372 by TwoRazorReign

Knowing that most people don't understand mathematics, I know how easy it is to mislead people by presenting mathematical formulas and big words. This is what is happening here and I am vehemently against it.


I understand you believe people are being mislead. This is unlikely. Most people reading this thread understand that the data presented by skimflux is just one of many interpretations of the effect Emerald Legionnaire on the tournament meta. A different interpretation was provided by the original poster, and they used personal experience instead of using topline descriptive data.

It is true that the reader must discern for themselves if what's being stated in this thread is hard evidence or just speculation. But, I believe everybody reading understands it's all speculation, leading to a fun discussion. I do not think people are seeing all the data presented by skimflux and saying, "Gee whiz, look at all those numbers! I can't understand what's being shown, so I'm just going to trust it's true."

I don't want black chantry to make any decisions on nerfs/bans/errata on false information.


Again, I don't think the people making decisions about this game are going to be convinced by anything in this thread, whether valid/invalid statistical analysis is included or not.

Start collecting ACTUALLY VALID data and then continue with this analysis. Start collecting decklists from all tournament participants so this analysis actually has some value.


Sure, but even if one were to do this, it's still a limited analysis of data (small sample size due to Emerald Legionnaire being relatively new). There's also no adjustment for player behavior (ie, whether the prevalence of Emerald Legionnaire is related it being a new card players want to try). In other words, what you are asking for would also still mislead people.

ps. I know you don't know what you are doing, because you wouldn't have gone this far had you realized the data is erroneous. If you want to practice your toolbox and collect course credit or whatever, you can do so, just try not to mislead readers of this site while doing it.


I think skimflux can post whatever he wants. This entire forum is all opinion and speculation. People just have to not believe everything that they read.
The following user(s) said Thank You: skimflux

Please Log in or Create an account to join the conversation.

More
04 Apr 2019 14:07 #94374 by DJHedgehog

Sure, but even if one were to do this, it's still a limited analysis of data (small sample size due to Emerald Legionnaire being relatively new). There's also no adjustment for player behavior (ie, whether the prevalence of Emerald Legionnaire is related it being a new card players want to try). In other words, what you are asking for would also still mislead people.


Here are the data points I would consider for a more statistically valid approach to analyzing the data.

-Every deck for the event (This is the largest and heaviest lift. I would require deck lists from players- but who wants to enter that data? Nobody. There is no real way to summarize a deck in a way that all organizers would understand, so anything other than full lists allow discretionary decisions and are questionable from a data standardization perspective).

-The corresponding GW/VP/Place of each deck, including finals seeding.

-Size of the event.

-Length of the event in rounds.

-The Final table seating, including special tie breakers to decide (we use die rolls, which doesn't really show if a "better" deck makes it to the finals and even have a chance to win).


I'm sure there is something I'm not thinking of but a lot of the questions I find myself asking when it comes to winning decks are answered above. Does 'x' deck win smaller events? Does it win shorter events? Did the winner benefit from finals seeding? How do the cards distribute from the lowest performers to the highest? What are the least used cards in the game, what are the most?

Please Log in or Create an account to join the conversation.

More
04 Apr 2019 14:36 #94376 by Mewcat
You also need to obtain the FUN RATE and if it falls outside standard deviation needs nerfing. People obsess with win rates and so on but there are plenty of other things to consider.

1) is it fun when some guy takes 1 action and gets 12 minions in play? does putting all this activity in the untap phase take away from the minion phase

2) EL or any minion spam (and multiact) leads to long turns and monopolizing of turn timer. when 1 guy takes 20 minute turns and the other 4 have 1 or 2 actions a turn this is snooze fest. similar for anything that shuffles deck.

3) is the card in itself ok but is it overly constraining the play space (limited viable deck choices)

win rates are ok but this shit isn't a job. Rock/paper/scissors may be perfectly balanced by win rates but the FUN RATE is lacking.

Please Log in or Create an account to join the conversation.

More
04 Apr 2019 15:43 #94377 by skimflux

The point I have been making (repeatedly) is that regardless of the methods you use (which may or may not be valid, its irrelevant), using them to analyze TWDA results in entirely useless, and worse: misleading data. In order to get meaningful data you must use data from ENTIRE tournaments.

<ranting over the same basic point>


I am quite aware of the limitations of the TWDA data. OTOH, it is what we have - it is pratically impossible to have full information on all tournament participating decks.
Even if you had that, you'd find the need to factor in the seatings for each round. And if you had that, you'd have to factor in the playing style of each player. And then maybe the current weather, and the phase of the moon , or whatever...

My point is that statistics can be misleading, but the alternative is not to do any analysis at all, and that just leads to people calling for bans on their pet hate du jour because "my friend won a bunch of games we played in the kitchen". Which we get too much of already...

The main achievement of statistical analysis is misleading politicians into making decisions they wouldn't have made if they actually knew what they were doing


Despite the popularity of vote decks in VTES I have no knowledge of any actual politician playing VTES or browsing this forum, so I think we're safe :lol:

Knowing that most people don't understand mathematics, I know how easy it is to mislead people by presenting mathematical formulas and big words.


My impression is that players of VTES are often math-inclined, some even proeficient professionals, and they mostly understand the limitations of statistics and the underlying assumptions of working with the TWDA data.

ps. I know you don't know what you are doing, because you wouldn't have gone this far had you realized the data is erroneous (....) just try not to mislead readers of this site while doing it.


AFAICT the TWDA data is not erroneous (those are real tournament winning decks), but it is incomplete and some assumptions must be made to extract useful information.
Because I do understand the limitations of the TWDA I take care not to jump to conclusions, and I took that stance in my posts - my assumptions and opinions are usually marked as such, feel free to ignore them.

Please Log in or Create an account to join the conversation.

More
Moderators: AnkhaKraus
Time to create page: 0.186 seconds
Powered by Kunena Forum