Dominion Strategy Forum

Please login or register.

Login with username, password and session length
Pages: [1] 2 3 ... 5  All

Author Topic: Need Help Developing a Ranking System  (Read 2261 times)

0 Members and 1 Guest are viewing this topic.

filovirus

  • Scout
  • ****
  • Offline Offline
  • Posts: 43
  • Respect: +30
    • View Profile
Need Help Developing a Ranking System
« on: September 22, 2017, 11:21:31 am »
0

I would like to put together a project that creates a tiered list of all the Dominion kingdom cards. This would be different than anything done in the past when considering card ranking.

It would be similar to list created by ehunt found here: http://forum.dominionstrategy.com/index.php?topic=13822.0, but with key difference. In ehunts list, he focused on key cards. I would rather have a focus more on the perceived cost to benefit ratio.

A few of problems with ehunts list. First, it was created in August, 2015, so it doesnt incorporate Adventures, Empires, the new cards added to Dominion: Base and Intrigue, a couple of Promos, nor the soon to be released Nocturne set. Second, it was created by one user, and adjusted by responses made in that specific thread, so it didnt incorporate a lot of community input. Third, it has three of the tiers in the common tiered system for ranking, but it is missing the crucial S tier.

This would also offer better card power understanding than currently found in the Qvist ranking system found here: http://forum.dominionstrategy.com/index.php?topic=16373.0. Although the Qvist ranking system is very thorough and informative, it too is missing some key aspects.

First, it does compare cards to each other that are of the same cost, but has no way of comparing cards to each other if they have different costs. So a Remodel can be ranked against Smithy, both of which are 4 cost cards, but cannot be ranked against Cellar, which is a 2 cost card. Second, it is very arbitrary. Is Smithy actually a better card overall than Remodel, or Poacher, or some other 4 cost card. In a recent article written by Adam Horton found here: http://adamhorton.com/flog/?p=538, his view of the card ranks differed by around 30% from the communities ranking. His point was that A bunch of different people rated Dominion cards using whatever criteria they felt was appropriate. I agree; this is a problem. Third, it is very hard to maintain. With each new expansion, a whole new poll needs to be completed incorporating new cards. This is both time consuming and tedious.

I would like to take a different approach than both ehunts and Qvists approach. I would like to create a tiered system that has 4 tiers. Each card would be assigned a tier, as voted on by the community. The 4 tiers would include an S tier, an A tier, a B, tier and a C tier.

The S tier would include those cards that are clearly overpowered. The ones whose cost to benefit ratio is very good. It would probably include ~3-7% of the cards overall. These would be the cards that cannot be ignored in any given kingdom. A prime example would be Chapel.

The A tier would be those cards that offer better than average cost to benefit ratio. It would probably consist of the upper 30% of total cards. Cards that fit well in most decks and synergize with many cards. I would rate Sentry as an A tier card.

The B tier would include those cards whose cost to benefit ratio of pretty accurate. It would probably consist of the middle 30% of total cards. Those cards that can fit well in a lot of decks, but may not be the main players all the time. They synergize well with some cards, but fall flat with others. Festival would be a prime example in my opinion.

The C tier would include those cards that are underperformers. It would probably include the bottom 30% of total cards. Those cards that may well synergize really well some few specific cards, but otherwise can usually be ignored in most kingdoms. Scout would easily fit in this category.

What I need is help defining the 4 groups. What is the wording to use in each of the four groups to most represent the cards found within each group? After the groups are defined, I would like to create a poll that the community can participate in that gives each card a tier designation. If we have enough participants, the cards can even be given a (-) or (+) to the tier they belong to.
Logged

faust

  • Torturer
  • *****
  • Offline Offline
  • Posts: 1703
  • Shuffle iT Username: faust
  • Respect: +2261
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #1 on: September 22, 2017, 11:51:13 am »
+3

I don't think "cost to effect" is a very good system for rating differently priced cards. What does it even mean? Is a $4 as good as a $2 in this system if it gives the benefit of 2 $2? If we do it like that, I can tell you right now that almost all $2s will be ranked C. Effect does not scale linearly with price.
Logged
Since the number of points is within a constant factor of the number of city quarters, in the long run we can get (4 - ε) ↑↑ n points in n turns for any ε > 0.

filovirus

  • Scout
  • ****
  • Offline Offline
  • Posts: 43
  • Respect: +30
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #2 on: September 22, 2017, 11:56:39 am »
0

I don't think "cost to effect" is a very good system for rating differently priced cards. What does it even mean? Is a $4 as good as a $2 in this system if it gives the benefit of 2 $2? If we do it like that, I can tell you right now that almost all $2s will be ranked C. Effect does not scale linearly with price.

You are correct. I would foresee that a lot of lower cost cards do indeed fall into the "C" category. In games like this, benefit should be increasing at a slightly faster rate than cost. This is what makes the higher costing cards worth it to spend coins on. However, I also foresee than many $2 cards do indeed fall in the "B" and "A" categories, with Chapel probably even falling into the "S" category.
Logged

werothegreat

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 7829
  • Shuffle iT Username: werothegreat
  • Prepare to be boarded!
  • Respect: +8864
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #3 on: September 22, 2017, 11:58:53 am »
+13

Logged
Contrary to popular belief, I do not run the wiki all on my own.  There are plenty of other people who are actively editing.  Go bother them!

Check out this fantasy epic adventure novel I wrote, the Broken Globe!  http://www.amazon.com/Broken-Globe-Tyr-Chronicles-Book-ebook/dp/B00LR1SZAS/

filovirus

  • Scout
  • ****
  • Offline Offline
  • Posts: 43
  • Respect: +30
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #4 on: September 22, 2017, 12:02:42 pm »
0



I understand what you're getting at. But most competitive games, esports, etc. are using this type of system to rank power levels. I don't find it crazy that Dominion could follow suit. It's already tried and true in other venues.
Logged

crj

  • Explorer
  • *****
  • Offline Offline
  • Posts: 339
  • Respect: +432
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #5 on: September 22, 2017, 12:09:20 pm »
+1

I understand what you're getting at. But most competitive games, esports, etc. are using this type of system to rank power levels. I don't find it crazy that Dominion could follow suit. It's already tried and true in other venues.
It would be pretty difficult to do for Dominion. Whatever structure you pick, exceptions will abound.

That's a large part of what gives Dominion its depth.
Logged

Qvist

  • Mountebank
  • *****
  • Online Online
  • Posts: 2348
  • Shuffle iT Username: Qvist
  • Respect: +3881
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #6 on: September 22, 2017, 12:12:07 pm »
+5

I can tell you the same thing that I told Adam and many other people. I am not sure why a broader rating/ranking system is better than a more in depth one. You would lose so much info.
Also btw, I allow comparing a $2 card to a $4 card. Many just don't do that as it is both tedious and way less relevant. I have a list for all the cards, I just didn't post it yet because I did got burned out and had another project going on. You could just take all the lists and put 80%+ cards in tier 1, 60-80% cards in tier 2, ... 0-20% in tier 5 and done. That is the big upside of using percentages and doing it the way I do it.
Logged

weesh

  • Alchemist
  • ***
  • Offline Offline
  • Posts: 35
  • MOAR MAGPIES
  • Respect: +73
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #7 on: September 22, 2017, 12:16:45 pm »
+1

Third, it has three of the tiers in the common tiered system for ranking, but it is missing the crucial S tier...

if you think 4 tiers are necessary, why not 'A' through 'D'?
Personally, I find 'S-tier' to be Stupid.

That said, I'm interested in seeing a list with the modern cards.
« Last Edit: September 22, 2017, 12:19:22 pm by weesh »
Logged

GendoIkari

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 6405
  • Respect: +7011
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #8 on: September 22, 2017, 12:28:18 pm »
+1

Third, it has three of the tiers in the common tiered system for ranking, but it is missing the crucial S tier...

if you think 4 tiers are necessary, why not 'A' through 'D'?
Personally, I find 'S-tier' to be Stupid.

That said, I'm interested in seeing a list with the modern cards.

It's not his invention... "S" tier is used in other tiered ranking systems such as the Smash Brothers tier lists. Dunno where it came from originally.
Logged
Check out my F.DS extension for Chrome! Card links; Dominion icons, and maybe more! http://forum.dominionstrategy.com/index.php?topic=13363.0

Thread for Firefox version:
http://forum.dominionstrategy.com/index.php?topic=16305.0

werothegreat

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 7829
  • Shuffle iT Username: werothegreat
  • Prepare to be boarded!
  • Respect: +8864
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #9 on: September 22, 2017, 12:48:18 pm »
+2

xkcd

I understand what you're getting at. But most competitive games, esports, etc. are using this type of system to rank power levels. I don't find it crazy that Dominion could follow suit. It's already tried and true in other venues.

But the thing is Dominion doesn't have a metagame the same way CCGs do.  Everything is dependent on what you're actually given.  Sure, Mountebank might be rated higher than Witch, but it doesn't matter if they're not both in the Kingdom.  We're never going to see something like "well I'll only ever put Mountebank in my deck, never Witch" because of the nature of the game.
Logged
Contrary to popular belief, I do not run the wiki all on my own.  There are plenty of other people who are actively editing.  Go bother them!

Check out this fantasy epic adventure novel I wrote, the Broken Globe!  http://www.amazon.com/Broken-Globe-Tyr-Chronicles-Book-ebook/dp/B00LR1SZAS/

filovirus

  • Scout
  • ****
  • Offline Offline
  • Posts: 43
  • Respect: +30
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #10 on: September 22, 2017, 12:51:39 pm »
0

I can tell you the same thing that I told Adam and many other people. I am not sure why a broader rating/ranking system is better than a more in depth one. You would lose so much info.
Also btw, I allow comparing a $2 card to a $4 card. Many just don't do that as it is both tedious and way less relevant. I have a list for all the cards, I just didn't post it yet because I did got burned out and had another project going on. You could just take all the lists and put 80%+ cards in tier 1, 60-80% cards in tier 2, ... 0-20% in tier 5 and done. That is the big upside of using percentages and doing it the way I do it.

Thanks for responding Qvist. I have watched all your YouTube videos on your 2016 ranking (you never completed the 5-cost or 6+cost cards). Your ranking system does offer a lot of good information. However, a few statistical items arise in your system of ranking. Firstly, the sample size is relatively small to get a truly representation of what the community thinks. With your system is would be tough to get a truly representative rank. Those that take your poll need to be very familiar with every card in the game and make non-arbitrary comparisons. The causal player with less understanding of each card will greatly skew your results. Second, there's bound to be some very high standard deviations for given cards because it is a mix of highly competitive and casual players, which would make too hard to put any credit in their ranking.

If I have a kingdom with Hamlet (ranked 5th for two cost cards), Wishing Well (ranked 26 for three cost cards), and Trader (ranked 54 for four cost cards), and have 4 coins in hand, which would be my best cost to benefit ratio? Hamlet would probably be "A" tier, Wishing Well "B" tier, and Trader "C" tier. Would I want to overpay for Hamlet? Overpay for Wishing Well? Or get a Trader? Obviously it may depend on the kingdom, but Hamlet may be the best option overall. It is tough to make those comparisons with such a stringent ranking system and not knowing where the 2-costers rank according to the higher costing other cards.

A 4 tier system would help alleviate the confusion.
Logged

filovirus

  • Scout
  • ****
  • Offline Offline
  • Posts: 43
  • Respect: +30
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #11 on: September 22, 2017, 12:58:28 pm »
0

xkcd

I understand what you're getting at. But most competitive games, esports, etc. are using this type of system to rank power levels. I don't find it crazy that Dominion could follow suit. It's already tried and true in other venues.

But the thing is Dominion doesn't have a metagame the same way CCGs do.  Everything is dependent on what you're actually given.  Sure, Mountebank might be rated higher than Witch, but it doesn't matter if they're not both in the Kingdom.  We're never going to see something like "well I'll only ever put Mountebank in my deck, never Witch" because of the nature of the game.

But it doesn't matter. Both would be probably be "A" tier cards. Then it's comparing the two within the other kingdom cards. But what about Junk Dealer compared to Festival? Both will probably have different tiers. Now the comparison would be valid. I also disagree that their is no META game in Dominion. Most Effective Tactic Available does show up in this game. Knowing Ill Gotten Gains is a strong META card, and Rabble not as much is part of Dominion's overall META. Whether they show up together in a kingdom or not, it is still worth while to know relative power.
Logged

trivialknot

  • Tactician
  • *****
  • Online Online
  • Posts: 437
  • Respect: +614
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #12 on: September 22, 2017, 01:06:17 pm »
+5

Thanks for responding Qvist. I have watched all your YouTube videos on your 2016 ranking (you never completed the 5-cost or 6+cost cards). Your ranking system does offer a lot of good information. However, a few statistical items arise in your system of ranking. Firstly, the sample size is relatively small to get a truly representation of what the community thinks. With your system is would be tough to get a truly representative rank. Those that take your poll need to be very familiar with every card in the game and make non-arbitrary comparisons. The causal player with less understanding of each card will greatly skew your results. Second, there's bound to be some very high standard deviations for given cards because it is a mix of highly competitive and casual players, which would make too hard to put any credit in their ranking.
(Emphasis mine)  You appear to be complaining that Qvist's sample is not representative of the community, and simultaneously too representative.

If I have a kingdom with Hamlet (ranked 5th for two cost cards), Wishing Well (ranked 26 for three cost cards), and Trader (ranked 54 for four cost cards), and have 4 coins in hand, which would be my best cost to benefit ratio? Hamlet would probably be "A" tier, Wishing Well "B" tier, and Trader "C" tier. Would I want to overpay for Hamlet? Overpay for Wishing Well? Or get a Trader? Obviously it may depend on the kingdom, but Hamlet may be the best option overall. It is tough to make those comparisons with such a stringent ranking system and not knowing where the 2-costers rank according to the higher costing other cards.
That... is exactly the wrong way to use a card ranking.
Logged

Qvist

  • Mountebank
  • *****
  • Online Online
  • Posts: 2348
  • Shuffle iT Username: Qvist
  • Respect: +3881
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #13 on: September 22, 2017, 01:08:07 pm »
+1


Thanks for responding Qvist. I have watched all your YouTube videos on your 2016 ranking (you never completed the 5-cost or 6+cost cards). Your ranking system does offer a lot of good information. However, a few statistical items arise in your system of ranking. Firstly, the sample size is relatively small to get a truly representation of what the community thinks. With your system is would be tough to get a truly representative rank. Those that take your poll need to be very familiar with every card in the game and make non-arbitrary comparisons. The causal player with less understanding of each card will greatly skew your results. Second, there's bound to be some very high standard deviations for given cards because it is a mix of highly competitive and casual players, which would make too hard to put any credit in their ranking.

That is why the voting is weighted. Higher rated player's votes have a higher weight than lower people's. And if you look at the history of past votings, it is pretty impressive to me how some cards stayed constistent over the time while others (where the meta indeed changed) have adjusted properly. I think the list are a reasonable well approximation at the strength. Having few voters is indeed a problem, but that is just due to the fact that this is Dominion and not Magic: The Gathering or Hearthstone. I can't change the popularity of the game.

If I have a kingdom with Hamlet (ranked 5th for two cost cards), Wishing Well (ranked 26 for three cost cards), and Trader (ranked 54 for four cost cards), and have 4 coins in hand, which would be my best cost to benefit ratio? Hamlet would probably be "A" tier, Wishing Well "B" tier, and Trader "C" tier. Would I want to overpay for Hamlet? Overpay for Wishing Well? Or get a Trader? Obviously it may depend on the kingdom, but Hamlet may be the best option overall. It is tough to make those comparisons with such a stringent ranking system and not knowing where the 2-costers rank according to the higher costing other cards.

A 4 tier system would help alleviate the confusion.

I don't understand how that would be better. The lists are not for telling you what you should buy if you have X coins in hand. You should buy what is best for your deck. If you need a Counting House, come and buy one. There is Chapel, better get all of them if it is the best card in Dominion. It doesn't tell you how many of each card you should get and in which order. That is not the purpose. It's just how impactful and relevant those cards are on the board. I think you are approaching those lists incorrectly.

Btw, I posted the rest of this year's list in the forum and it is on the wiki as well.
Logged

markusin

  • Cartographer
  • *****
  • Offline Offline
  • Posts: 3267
  • Shuffle iT Username: markusin
  • I also switched from Starcraft
  • Respect: +1876
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #14 on: September 22, 2017, 01:24:11 pm »
0

I can tell you the same thing that I told Adam and many other people. I am not sure why a broader rating/ranking system is better than a more in depth one. You would lose so much info.
Also btw, I allow comparing a $2 card to a $4 card. Many just don't do that as it is both tedious and way less relevant. I have a list for all the cards, I just didn't post it yet because I did got burned out and had another project going on. You could just take all the lists and put 80%+ cards in tier 1, 60-80% cards in tier 2, ... 0-20% in tier 5 and done. That is the big upside of using percentages and doing it the way I do it.

Thanks for responding Qvist. I have watched all your YouTube videos on your 2016 ranking (you never completed the 5-cost or 6+cost cards). Your ranking system does offer a lot of good information. However, a few statistical items arise in your system of ranking. Firstly, the sample size is relatively small to get a truly representation of what the community thinks. With your system is would be tough to get a truly representative rank. Those that take your poll need to be very familiar with every card in the game and make non-arbitrary comparisons. The causal player with less understanding of each card will greatly skew your results. Second, there's bound to be some very high standard deviations for given cards because it is a mix of highly competitive and casual players, which would make too hard to put any credit in their ranking.

If I have a kingdom with Hamlet (ranked 5th for two cost cards), Wishing Well (ranked 26 for three cost cards), and Trader (ranked 54 for four cost cards), and have 4 coins in hand, which would be my best cost to benefit ratio? Hamlet would probably be "A" tier, Wishing Well "B" tier, and Trader "C" tier. Would I want to overpay for Hamlet? Overpay for Wishing Well? Or get a Trader? Obviously it may depend on the kingdom, but Hamlet may be the best option overall. It is tough to make those comparisons with such a stringent ranking system and not knowing where the 2-costers rank according to the higher costing other cards.

A 4 tier system would help alleviate the confusion.

This doesn't tie in so well with Tier being determined by effect/cost. Wishing Well might be B tier and less efficient than the A tier Hamlet, but Wishing Well costs $3 instead of $2 and maybe to total effect of Wishing Well is higher than Hamlet. This becomes more likely with the B tier cards that cost $4, and even more likely with the $5 cost B tiers and so on.

To know what to buy with $4 using tiers, with cards cheaper than $4 also being considered, the tier of each card would have to be determined considering only its effect, ignoring cost altogether.

Qvist's ranking doesn't ignore card cost when factoring rank, but neither does the proposed tier system. Whether you use Qvist ranking or tiers, solving the problem of knowing what to buy requires cards to be compared to all other cards in terms of raw strength, ignoring cost.

This is what is at the heart of Adam's criticism of variations in how players rank cards making the ranking system not as helpful as it could be. Regardless of ranking scheme, what should be emphasized is how players are expected to compare cards.
Logged

Seprix

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5291
  • Shuffle iT Username: Seprix
  • Respect: +2976
    • View Profile
    • The Border Village
Re: Need Help Developing a Ranking System
« Reply #15 on: September 22, 2017, 02:36:24 pm »
+3

This is a terrible idea, just drop it altogether. Ratio of cost to effectiveness doesn't help anyone play any board better. We have Qvist rankings and now Adam rankings, do we need a third list? Just take Adam's rankings and tier those.
Logged
the weed whackers will return

filovirus

  • Scout
  • ****
  • Offline Offline
  • Posts: 43
  • Respect: +30
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #16 on: September 22, 2017, 06:15:30 pm »
0

This is a terrible idea, just drop it altogether. Ratio of cost to effectiveness doesn't help anyone play any board better. We have Qvist rankings and now Adam rankings, do we need a third list? Just take Adam's rankings and tier those.

Thanks for your opinion. I tend to think that the Adam rankings are probably a little more accurate than the Qvist rankings, but the problem that keeps coming back is that it is based on 1 person's assessment of each individual card, as opposed to a community poll. The Qvist rankings can be a better indication of community assessment IF a sample size were larger and ranking guidelines were given OR if the poll was only given to a select few who are considered the "best" at the game. This would alleviate the high standard deviations that some cards received.

I think I will still try to create a poll. I will probably put it on boardgamegeek, just because the polling system is much more advanced on that site than this site. Am I allowed to create a thread with a link to that site? If there is poor turnout or as much kickback there as here, I can eventually scrap the whole concept.
Logged

filovirus

  • Scout
  • ****
  • Offline Offline
  • Posts: 43
  • Respect: +30
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #17 on: September 22, 2017, 06:26:51 pm »
0

I don't understand how that would be better. The lists are not for telling you what you should buy if you have X coins in hand. You should buy what is best for your deck. If you need a Counting House, come and buy one. There is Chapel, better get all of them if it is the best card in Dominion. It doesn't tell you how many of each card you should get and in which order. That is not the purpose. It's just how impactful and relevant those cards are on the board. I think you are approaching those lists incorrectly.

Btw, I posted the rest of this year's list in the forum and it is on the wiki as well.

I never implied that it would tell you which card to buy. It would only give an indication of the relative power and usefulness of the card. Nor did I imply that it would mean to purchase all ten Chapels because it is a good card. My poor wording probably led to your misunderstanding.

What I would like, is a quick system where a player such as me, who has no experience with Empires, Cornucopia, Guilds, or Dark Ages, can evaluate quickly the given cards in a kingdom and what the community thinks of their usefulness. Is Chariot Race good in most decks? I don't have the foggiest idea. I think this idea could help out those casual players as well.

Of course it will have no usefulness for those experienced competitive players. But then again, neither do the Qvist rankings or Adam's rankings.

So as a returning player to Dominion, this is what I would like to see.

And Qvist, truly I like your ranking system. It is very informative. Sorry if I came across as completely dismissive to it.
Logged

filovirus

  • Scout
  • ****
  • Offline Offline
  • Posts: 43
  • Respect: +30
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #18 on: September 22, 2017, 06:32:24 pm »
0

These are the tier ranking qualifiers I came up. I would like some critique. Also the Dominion: Base cards as I would order them according to the qualifiers.

S This card is good in 95% of decks. This card makes 95% of decks better. This card synergizes well with most other cards. This card cannot be ignored when in any given kingdom. If my opponent gets this card and I dont, there is a good chance he will win. If I get this card and my opponent doesnt, there is a good chance I will win. This card is the focal point of a major strategy.

A This card is good in 75% or more of decks. This card makes 95% of decks better. This card synergizes well with most other cards. This card should not be ignored when in any given kingdom. If my opponent gets this card, and I dont, he is likely to win, but other strategies may be just as good. This card can be a focal point of a major strategy OR this card is a good support card for major strategy.

B This card is good in 50% or more of decks. This card makes 75% of decks better. This card synergizes well with about 50% of other cards. This card can be ignored when in any given kingdom. If my opponent gets this card, and I dont, the outcome is still based on who plays better and not on card interaction.  This card is rarely a focal point of a major strategy, but often a good support card for a major strategy.

C This card is good in less than 50% of decks. This card makes 50% or decks better. It is difficult at times to get good use from this card. This card often is ignored when in any given kingdom. If my opponent gets this card, and I dont, the outcome is still based on who plays better and not on card interaction. This card is never a focal point of a major strategy, and rarely a good support card. However, this card may have amazing synergy with another, specific card.

Celler - B
Chapel - S
Moat - B
Harbinger - C
Merchant - B
Vassal - C
Village - B
Workshop - C
Bureaucrat - B
Gardens - C
Militia - A
Moneylender - B
Poacher - A
Remodel - B
Smithy - A
Throne Room A
Bandit - B
Council Room - A
Festival - B
Laboratory - A
Library - C
Market - A
Mine - C
Sentry - A
Witch - S
Artisan - B
Logged

Mic Qsenoch

  • 2015 DS Champion
  • *
  • Offline Offline
  • Posts: 1586
  • Respect: +3943
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #19 on: September 22, 2017, 07:06:59 pm »
+4

Cellar - C
Chapel - S
Moat - B
Harbinger - B
Merchant - B
Vassal - C
Village - A
Workshop - B
Bureaucrat - D
Gardens - D
Militia - A
Moneylender - A
Poacher - B
Remodel - B
Smithy - A
Throne Room S
Bandit - B
Council Room - B
Festival - A
Laboratory - A
Library - B
Market - B
Mine - D
Sentry - S
Witch - S
Artisan - S
« Last Edit: September 22, 2017, 07:08:37 pm by Mic Qsenoch »
Logged

filovirus

  • Scout
  • ****
  • Offline Offline
  • Posts: 43
  • Respect: +30
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #20 on: September 22, 2017, 07:26:47 pm »
0

Cellar - C
Chapel - S
Moat - B
Harbinger - B
Merchant - B
Vassal - C
Village - A
Workshop - B
Bureaucrat - D
Gardens - D
Militia - A
Moneylender - A
Poacher - B
Remodel - B
Smithy - A
Throne Room S
Bandit - B
Council Room - B
Festival - A
Laboratory - A
Library - B
Market - B
Mine - D
Sentry - S
Witch - S
Artisan - S

Thanks. As you can see, according my criteria for each tier, you have far too many "S" cards. "S" should be saved for only the best of the best. This is what I need help with. Creating a set of criteria that produces an even distribution for A, B, and C, with few cards ever reaching the "S" tier.
Logged

Mic Qsenoch

  • 2015 DS Champion
  • *
  • Offline Offline
  • Posts: 1586
  • Respect: +3943
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #21 on: September 22, 2017, 07:57:25 pm »
+6

In my opinion, rating is much more valuable than ranking, no need to shoehorn things into distributions or worry about precise orderings.
Logged

trivialknot

  • Tactician
  • *****
  • Online Online
  • Posts: 437
  • Respect: +614
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #22 on: September 22, 2017, 08:55:12 pm »
+3

Thanks for your opinion. I tend to think that the Adam rankings are probably a little more accurate than the Qvist rankings, but the problem that keeps coming back is that it is based on 1 person's assessment of each individual card, as opposed to a community poll.
Adam's ranking wasn't just his own assessment.  It was based on a poll of 35 people.

It's nice that Adam performed an independent poll, but I think the most useful contribution he made was not the data itself, but the analysis and discussion.  If he had instead borrowed data from Qvist and analyzed that, it might have been slightly better or slightly worse, but it wouldn't have made a huge difference.  Like, he still would have been able to write a pretty cool article and all.

I am skeptical of the value of giving survey respondents very specific instructions on how to rate cards.  The thing is, you don't actually know that people are following your instructions.  Imagine all the people in this thread who are disagreeing with you, but now instead of voicing their disagreement out loud, they are quietly filling out your survey according to their own views.  It just can't be helped.
Logged

DG

  • Cartographer
  • *****
  • Offline Offline
  • Posts: 3887
  • Respect: +2403
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #23 on: September 22, 2017, 09:07:39 pm »
+4

These sorts of ranking systems will only be any good if they can rate cards like outpost. There tends to be one set of kingdoms where outpost is useless and another set of kingdoms where outpost is fantastic and not much in between. You usually don't need to buy many outposts even when they're good and they are not going to score you any points on their own. The pile is one of the least likely to run out. It's still however a key card and when outpost is good, a person who buys outpost will reliably beat a person who doesn't.

There's not much point in creating a new set of rankings that can't rate cards like outpost.
Logged

Beyond Awesome

  • Global Moderator
  • *****
  • Offline Offline
  • Posts: 2893
  • Shuffle iT Username: Beyond Awesome
  • Respect: +2379
    • View Profile
Re: Need Help Developing a Ranking System
« Reply #24 on: September 22, 2017, 09:26:55 pm »
+1

You shouldn't rank on how often you buy, but probably impact instead or a combination of the two.
Logged
Pages: [1] 2 3 ... 5  All
 

Page created in 0.225 seconds with 20 queries.