Post by larryturbo on Jan 2, 2021 23:14:13 GMT -5
First time poster long time Scooby-Doo fan. I just wanted to write about my newfound love for this Scooby-Doo card game from my childhood.
Back in 2004 Deagostini released a series of Scooby Doo magazines. The series consisted of 101 Issues and lasted for 2 years, with one new issue every week. Included with each Issue was a pack of 8 cards plus one extra Puzzle card every 3rd issue.
A grand total of 375 cards were released of which 275 were “Location” cards. Those location cards were further divided into sets named after continents. There were 77 cards in the Europe set, 60 in the Asia, 52 in the Americas, 48 in the Africa, 36 in the Oceania and 2 special Antarctica cards released one every Christmas.
Each card featured either a member of the gang or a villain in a famous location. They also featured 4 stats (Size, Age, Beauty and Spookiness) that could be used to “do battle” using the Top Trumps rules.
Pictures of all those can be found in the trading card database here:
www.tcdb.com/ViewSet.cfm/sid/125883/2004-DeAgostini-Scooby-Doo-World-of-Mystery
But that is not what this thread is about. What I wanted was to make a tier list in excel and see which cards were the best and which were the worst. So for the lovers of data and Scooby doo this will be a pleasant read.
Starting with the general character distribution in all the Location cards. To get you in the mood for some numbers.
Then we can break it down per set.
Plus the 2 cheeky Velma Antarctica cards. (Which may or may not have been Greece exclusive but they count in my book).
For the purpose of organization. I call the Europe cards “E-#”, the Asia “S-#”, the Americas “M-#”, the Africa “F-#”, the Oceania “O-#” and the Antarctica “N-#”. So Card E-33 is card “Shaggy at the Tower of Belem” 33/77 from the Europe set.
Now, how can we rank those cards in a meaningful manner? Well, each card has a 4 stats which range from 1 to 100. So a card can have a minimum of 4 and a maximum of 400 stat total. A first thought would be to find what cards have the biggest overall numbers.
We get the Top 10 table below.
From this we can say that card E-46 “Daphne at the Northern Lights” is the best card in the game. With an over 25 point margin from the next best card “Fred in the Andes”. It is almost perfect. But since each duel only happens between one stat at a time. It is also useful to have the best cards for each stat.
They are as follows:
Size
Age
Beauty
Spookiness
In the same vain we can find the worst cards in the game. Both for overall stats and for individual ones.
And Individually:
Size
Age
Beauty
Spookiness
It is also possible to find the most average card. Since all the numbers are random 1-100, the average stat should be 50.5 and the average total four times that at 202. However, since this isn’t an infinite sample size, we can plot the stats into bell curves to find the actual average in our case.
The average total stat in our data is 216.
The average size is 52.
The average age is 51.
The average beauty is 53.
The average spookiness is 60. A statistically significant difference from the expected 50-51. Maybe some cards were manually made more spooky or it might really be random.
Thus card S-2 “Fred at Varanasi” is the most average card of them all. Being the closest to the peak on all 5 curves.
And now comes the question. Is this enough info to rank the cards and make the best deck?
The answer is No.
Buckle your seatbelts because we are about to go deeper into the rabbit-hole. This game is played with 2 or more player each taking turns, reading one of the stats on the card at the top of their deck, and then comparing whose number is biggest. The biggest number wins.
Let’s now imagine 2 hypothetical cards.
Same stat total but which card is more useful? The first card is great to see on your turn, since you will always choose the 100 and win almost every duel, (almost because a 100 can’t win another 100). The second card might be better to have when your opponent is playing and you need something more defensive.
This leads us to define a new term, which is the Usefulness of the card.
For every stat on one card there are 275 possible duels with the same stat of all the other cards. For 4 stats that is 254 x 4 = 1100 possible duels for every card in the game. The best a card could do (if it had 100-100-100-100 and no other card had a single 100) would be to win 1006 of those duels. (Those last 4 are with itself and those will always be ties.)
Now if we calculate how many of the 1100 duels each card can win we will have a metric for the true Usefulness of each card. The more duels a card can win the better it is both offensively and defensively. And we can express that in percentages in the form of (winning duels/1100) *100%.
We get the following Top and Bottom 10s.
The tables are similar but not the same. Cards F-29 and F-1 were top 3 and 4 in the stats table sharing a 335 stat total. But with this usefulness stat we can tell that F-29 is 1% better than F-1. And look at Card M-3, it only just made the Top 10 in the stat total but its spread is so good that it could catapult to 6th best in true usefulness.
On the ugly side, we get another confirmation that card F-38 is not just bad, it is abysmal, with a 3% win-rate. 8% lower than the second worst. (It is thus recommended that this card is played as (30-6-19-12) instead of (3-6-19-12) just to make it less bad. It would still be the worst but with a 10% win-rate this time.)
The most average card now becomes:
Being able to win against 50% all the duels.
Now we have all we need to make a Tier list.
Each tier differing by 10% from the next. And it looks like this:
Technically we have exhausted all competitive data and we know exactly which cards should be in a deck and which cards shouldn't, but we can also find out some other things.
Like which character is best.
That would be Fred. And despite Daphne having the best card in the game she has the worst performing cards of all characters.
I hear you ask: Why would anyone want to know all this stuff about this game that no one has played since 2004? All I can say is that little Johnny had a very hard time beating dad this Christmas. He had to train a lot with mom and her mysteriously easier deck, working to figure out which cards were good.
Back in 2004 Deagostini released a series of Scooby Doo magazines. The series consisted of 101 Issues and lasted for 2 years, with one new issue every week. Included with each Issue was a pack of 8 cards plus one extra Puzzle card every 3rd issue.
A grand total of 375 cards were released of which 275 were “Location” cards. Those location cards were further divided into sets named after continents. There were 77 cards in the Europe set, 60 in the Asia, 52 in the Americas, 48 in the Africa, 36 in the Oceania and 2 special Antarctica cards released one every Christmas.
Each card featured either a member of the gang or a villain in a famous location. They also featured 4 stats (Size, Age, Beauty and Spookiness) that could be used to “do battle” using the Top Trumps rules.
Pictures of all those can be found in the trading card database here:
www.tcdb.com/ViewSet.cfm/sid/125883/2004-DeAgostini-Scooby-Doo-World-of-Mystery
But that is not what this thread is about. What I wanted was to make a tier list in excel and see which cards were the best and which were the worst. So for the lovers of data and Scooby doo this will be a pleasant read.
Starting with the general character distribution in all the Location cards. To get you in the mood for some numbers.
Then we can break it down per set.
Plus the 2 cheeky Velma Antarctica cards. (Which may or may not have been Greece exclusive but they count in my book).
For the purpose of organization. I call the Europe cards “E-#”, the Asia “S-#”, the Americas “M-#”, the Africa “F-#”, the Oceania “O-#” and the Antarctica “N-#”. So Card E-33 is card “Shaggy at the Tower of Belem” 33/77 from the Europe set.
Now, how can we rank those cards in a meaningful manner? Well, each card has a 4 stats which range from 1 to 100. So a card can have a minimum of 4 and a maximum of 400 stat total. A first thought would be to find what cards have the biggest overall numbers.
We get the Top 10 table below.
From this we can say that card E-46 “Daphne at the Northern Lights” is the best card in the game. With an over 25 point margin from the next best card “Fred in the Andes”. It is almost perfect. But since each duel only happens between one stat at a time. It is also useful to have the best cards for each stat.
They are as follows:
Size
Age
Beauty
Spookiness
In the same vain we can find the worst cards in the game. Both for overall stats and for individual ones.
And Individually:
Size
Age
Beauty
Spookiness
It is also possible to find the most average card. Since all the numbers are random 1-100, the average stat should be 50.5 and the average total four times that at 202. However, since this isn’t an infinite sample size, we can plot the stats into bell curves to find the actual average in our case.
The average total stat in our data is 216.
The average size is 52.
The average age is 51.
The average beauty is 53.
The average spookiness is 60. A statistically significant difference from the expected 50-51. Maybe some cards were manually made more spooky or it might really be random.
Thus card S-2 “Fred at Varanasi” is the most average card of them all. Being the closest to the peak on all 5 curves.
And now comes the question. Is this enough info to rank the cards and make the best deck?
The answer is No.
Buckle your seatbelts because we are about to go deeper into the rabbit-hole. This game is played with 2 or more player each taking turns, reading one of the stats on the card at the top of their deck, and then comparing whose number is biggest. The biggest number wins.
Let’s now imagine 2 hypothetical cards.
- Card #1 with stats 100-1-1-1 with a stat total of 103.
- Card #2 with stats 25-26-26-26 again with a stat total of 103.
Same stat total but which card is more useful? The first card is great to see on your turn, since you will always choose the 100 and win almost every duel, (almost because a 100 can’t win another 100). The second card might be better to have when your opponent is playing and you need something more defensive.
This leads us to define a new term, which is the Usefulness of the card.
For every stat on one card there are 275 possible duels with the same stat of all the other cards. For 4 stats that is 254 x 4 = 1100 possible duels for every card in the game. The best a card could do (if it had 100-100-100-100 and no other card had a single 100) would be to win 1006 of those duels. (Those last 4 are with itself and those will always be ties.)
Now if we calculate how many of the 1100 duels each card can win we will have a metric for the true Usefulness of each card. The more duels a card can win the better it is both offensively and defensively. And we can express that in percentages in the form of (winning duels/1100) *100%.
We get the following Top and Bottom 10s.
The tables are similar but not the same. Cards F-29 and F-1 were top 3 and 4 in the stats table sharing a 335 stat total. But with this usefulness stat we can tell that F-29 is 1% better than F-1. And look at Card M-3, it only just made the Top 10 in the stat total but its spread is so good that it could catapult to 6th best in true usefulness.
On the ugly side, we get another confirmation that card F-38 is not just bad, it is abysmal, with a 3% win-rate. 8% lower than the second worst. (It is thus recommended that this card is played as (30-6-19-12) instead of (3-6-19-12) just to make it less bad. It would still be the worst but with a 10% win-rate this time.)
The most average card now becomes:
Being able to win against 50% all the duels.
Now we have all we need to make a Tier list.
Each tier differing by 10% from the next. And it looks like this:
Technically we have exhausted all competitive data and we know exactly which cards should be in a deck and which cards shouldn't, but we can also find out some other things.
Like which character is best.
That would be Fred. And despite Daphne having the best card in the game she has the worst performing cards of all characters.
I hear you ask: Why would anyone want to know all this stuff about this game that no one has played since 2004? All I can say is that little Johnny had a very hard time beating dad this Christmas. He had to train a lot with mom and her mysteriously easier deck, working to figure out which cards were good.