Statistics are gotten from here.
http://cap.smogon.com/ShoddyStats/2009-04/Smogon-200904-Details-Standard.html
I used my calculator here.
http://www.smogon.com/forums/showthread.php?t=54466 .
Sometimes when I refer to pokemon, I am refering to the metagame. Think of it as a fictional pokemon who is the average of the metagame.
First I went through the list and got the usage PERCENT for each pokemon. Any pokemon who had a usage percent of 0 were excluded. This statistic is now known as raw_pk_usage_perc.
I then corrected the statistic using the following.
cor_pk_usage_perc = raw_pk_usage_perc / sum of all(raw_pk_usage_perc)
I'm not going to post equations all over the place, as that will bore me. Assume whenever I correct a usage percentage I use an equation similar to the one above.
Top 5 Pokemon Used.
Scizor: 4.6%
Salamence: 3.9%
Gyarados 3.4%
Heatran 3.3%
Infernape: 3.0%
This means if you were to battle on smogon in April and your opponent sent out a random pokemon. It has a 4.6% chance of being Scizor. Obviously is Scizor has been fainted it can't be anymore, but I need these numbers specifically for what I am going to do later.
-----------------------
Next I went back to the raw stats and got the usage percent for each of the moves. I then corrected this as I did with the pokemon usage numbers..
Top 5 Moves Used.
Earthquake 6.2%
Thunderbolt 3.5%
Stealth Rock 3.2%
Stone Edge 3.0%
Surf 2.5%
I made another statistic that is the corrected move percents for only attacks by excluding any “Other” type moves. I'll call it the cor_attack_perc.
Top 5 ATTACKS Used.
Earthquake 9.1%
Thunderbolt 5.1%
Stone Edge 4.5%
Surf 3.7%
Ice Beam 3.6%
I made another statistic that by adding up the corrected percentages of attacks that are of a type.
Percent Usage for each Attack Type
Ground 10.8%
Fighting 10.1%
Ice 9.6%
Fire 9.2%
Normal 8.5%
Water 7.4%
Electric 7.3%
Dark 5.8%
Steel 5.1%
Dragon 5.0%
Bug 4.8%
Rock 4.8%
Grass 4.2%
Ghost 2.9%
Flying 2.2%
Psychic 2.0%
Poison 0.3%
This means that if an attack is used on shoddy in april at random, It will be a ground type attack 10.8% of the time and a poison type attack 0.3% of the time.
----------------------------------
I made a list with all these pokemon and their usage percentages. Each pokemon has a weakness that can be expressed as a number (0.25, 0.5, 1, 2, or 4). From this I was able to calculate the average weakness of the April metagame.
average_weakness_type = sum of all ( pokemons weakness to type * corrected usage percent)
Again, I'm not going to keep posting rehashing of equations you have basically seen. Assume I use this equation to get all averages. I may have used this equation above somewhere. Forgive me as I have been rearranging this post.
The Average Weakness of Metagame
Fire: 1.238,
Ice: 1.214,
Ground: 1.208,
Rock: 1.109,
Fighting: 1.109,
Electric: 1.084,
Water: 1.066,
Flying: 1.052,
Dark: 1.026,
Ghost: 0.987,
Dragon: 0.986,
Grass: 0.965,
Bug: 0.951,
Psychic: 0.900,
Steel: 0.825,
Normal: 0.795,
Poison: 0.706
That means that if you were to battle a training on shoddy in april. The pokemon that they sent out will, on average, take 1.238 times damage from fire attacks. and 0.706 times damage from poison attacks.
Probable contributers to the fire and poison stats are steel types. Steel takes 0 from poison and 2 from fire.
-----------------------------
I then ASSUMED (you can argue whether I can or cannot, I don't care) that the Average Weakness of the Metagame should be equal to the attack type distribution. This is hard to characterize since it's a bit of an abstract example. It's a bit inaccurate of a metric but it gives a good starting place. The best way to explain this is to look at an extreme. If pokemon take 0* damage from normal type attacks, then there should be 0 normal type attacks in the metagame. The other extreme doesn't work out since if a pokemon takes 4* damage from fire type attacks, then there shouldn't be any other type of move but fire. If somebody could come up with a corrected formula be my guest.
Percent of Weaknesses Weak to Type
Fire 7.2%
Ice 7.1%
Ground 7.0%
Rock 6.4%
Fighting 6.4%
Electric 6.3%
Water 6.2%
Flying 6.1%
Dark 6.0%
Ghost 5.7%
Dragon 5.7%
Grass 5.6%
Bug 5.5%
Psychic 5.2%
Steel 4.8%
Normal 4.6%
Poison 4.1%
I then compared the actual attack usage to the predicted attack usage based on pokemon weaknesses to get a measure of how overused a attack type is.
overused_attack_type_metric = actual_usage_percentage / predicted_usage_percentage
This will tell me how overused a attack is. Numbers above 1 indicate that the attack type is overused. Numbers below 1 indicate that the attack type is under used.
Overused Attack Type Metric
Normal 1.84,
Fighting 1.57,
Ground 1.54,
Ice 1.36,
Fire 1.28,
Water 1.20,
Electric 1.16,
Steel 1.07,
Dark 0.97,
Dragon 0.87,
Bug 0.87,
Grass 0.74,
Rock 0.74,
Ghost 0.51,
Psychic 0.38,
Flying 0.36,
Poison 0.08
So normal type moves are overused the most and poison type moves are underused the most.
I'll post more statistics as I get them.
Some Random Statistics
60% of all attacks were physical
40% of all attacks were special
The average power for an attack is 90.85
The average accuracy for an attack is 96.61
The average pp for an attack is 13.14
The Average Power of Attacks by Type
Normal 134.5,
Dragon 113.8,
Fire 101.0,
Fighting 100.4,
Rock 99.1,
Ground 98.3,
Water 90.8,
Electric 88.9,
Psychic 85.6,
Poison 85.1,
Flying 84.7,
Ice 79.2,
Bug 77.8,
Ghost 75.1,
Dark 61.3,
Grass 52.5,
Steel 51.4
(Note: Gyro Ball is counted as having 0 BP, which is what contributes to Steel being so Low. Grass knot is also counted as 0 BP)
Interesting questions.
Why are pokemon so weak to flying type moves and why are they so overused?
Does poison suck THAT badly?
Is the reason physical attacks used over special attacks because special attacks can be walled harder (I'm looking at you blissey)?
http://cap.smogon.com/ShoddyStats/2009-04/Smogon-200904-Details-Standard.html
I used my calculator here.
http://www.smogon.com/forums/showthread.php?t=54466 .
Sometimes when I refer to pokemon, I am refering to the metagame. Think of it as a fictional pokemon who is the average of the metagame.
First I went through the list and got the usage PERCENT for each pokemon. Any pokemon who had a usage percent of 0 were excluded. This statistic is now known as raw_pk_usage_perc.
I then corrected the statistic using the following.
cor_pk_usage_perc = raw_pk_usage_perc / sum of all(raw_pk_usage_perc)
I'm not going to post equations all over the place, as that will bore me. Assume whenever I correct a usage percentage I use an equation similar to the one above.
Top 5 Pokemon Used.
Scizor: 4.6%
Salamence: 3.9%
Gyarados 3.4%
Heatran 3.3%
Infernape: 3.0%
This means if you were to battle on smogon in April and your opponent sent out a random pokemon. It has a 4.6% chance of being Scizor. Obviously is Scizor has been fainted it can't be anymore, but I need these numbers specifically for what I am going to do later.
-----------------------
Next I went back to the raw stats and got the usage percent for each of the moves. I then corrected this as I did with the pokemon usage numbers..
Top 5 Moves Used.
Earthquake 6.2%
Thunderbolt 3.5%
Stealth Rock 3.2%
Stone Edge 3.0%
Surf 2.5%
I made another statistic that is the corrected move percents for only attacks by excluding any “Other” type moves. I'll call it the cor_attack_perc.
Top 5 ATTACKS Used.
Earthquake 9.1%
Thunderbolt 5.1%
Stone Edge 4.5%
Surf 3.7%
Ice Beam 3.6%
I made another statistic that by adding up the corrected percentages of attacks that are of a type.
Percent Usage for each Attack Type
Ground 10.8%
Fighting 10.1%
Ice 9.6%
Fire 9.2%
Normal 8.5%
Water 7.4%
Electric 7.3%
Dark 5.8%
Steel 5.1%
Dragon 5.0%
Bug 4.8%
Rock 4.8%
Grass 4.2%
Ghost 2.9%
Flying 2.2%
Psychic 2.0%
Poison 0.3%
This means that if an attack is used on shoddy in april at random, It will be a ground type attack 10.8% of the time and a poison type attack 0.3% of the time.
----------------------------------
I made a list with all these pokemon and their usage percentages. Each pokemon has a weakness that can be expressed as a number (0.25, 0.5, 1, 2, or 4). From this I was able to calculate the average weakness of the April metagame.
average_weakness_type = sum of all ( pokemons weakness to type * corrected usage percent)
Again, I'm not going to keep posting rehashing of equations you have basically seen. Assume I use this equation to get all averages. I may have used this equation above somewhere. Forgive me as I have been rearranging this post.
The Average Weakness of Metagame
Fire: 1.238,
Ice: 1.214,
Ground: 1.208,
Rock: 1.109,
Fighting: 1.109,
Electric: 1.084,
Water: 1.066,
Flying: 1.052,
Dark: 1.026,
Ghost: 0.987,
Dragon: 0.986,
Grass: 0.965,
Bug: 0.951,
Psychic: 0.900,
Steel: 0.825,
Normal: 0.795,
Poison: 0.706
That means that if you were to battle a training on shoddy in april. The pokemon that they sent out will, on average, take 1.238 times damage from fire attacks. and 0.706 times damage from poison attacks.
Probable contributers to the fire and poison stats are steel types. Steel takes 0 from poison and 2 from fire.
-----------------------------
I then ASSUMED (you can argue whether I can or cannot, I don't care) that the Average Weakness of the Metagame should be equal to the attack type distribution. This is hard to characterize since it's a bit of an abstract example. It's a bit inaccurate of a metric but it gives a good starting place. The best way to explain this is to look at an extreme. If pokemon take 0* damage from normal type attacks, then there should be 0 normal type attacks in the metagame. The other extreme doesn't work out since if a pokemon takes 4* damage from fire type attacks, then there shouldn't be any other type of move but fire. If somebody could come up with a corrected formula be my guest.
Percent of Weaknesses Weak to Type
Fire 7.2%
Ice 7.1%
Ground 7.0%
Rock 6.4%
Fighting 6.4%
Electric 6.3%
Water 6.2%
Flying 6.1%
Dark 6.0%
Ghost 5.7%
Dragon 5.7%
Grass 5.6%
Bug 5.5%
Psychic 5.2%
Steel 4.8%
Normal 4.6%
Poison 4.1%
I then compared the actual attack usage to the predicted attack usage based on pokemon weaknesses to get a measure of how overused a attack type is.
overused_attack_type_metric = actual_usage_percentage / predicted_usage_percentage
This will tell me how overused a attack is. Numbers above 1 indicate that the attack type is overused. Numbers below 1 indicate that the attack type is under used.
Overused Attack Type Metric
Normal 1.84,
Fighting 1.57,
Ground 1.54,
Ice 1.36,
Fire 1.28,
Water 1.20,
Electric 1.16,
Steel 1.07,
Dark 0.97,
Dragon 0.87,
Bug 0.87,
Grass 0.74,
Rock 0.74,
Ghost 0.51,
Psychic 0.38,
Flying 0.36,
Poison 0.08
So normal type moves are overused the most and poison type moves are underused the most.
I'll post more statistics as I get them.
Some Random Statistics
60% of all attacks were physical
40% of all attacks were special
The average power for an attack is 90.85
The average accuracy for an attack is 96.61
The average pp for an attack is 13.14
The Average Power of Attacks by Type
Normal 134.5,
Dragon 113.8,
Fire 101.0,
Fighting 100.4,
Rock 99.1,
Ground 98.3,
Water 90.8,
Electric 88.9,
Psychic 85.6,
Poison 85.1,
Flying 84.7,
Ice 79.2,
Bug 77.8,
Ghost 75.1,
Dark 61.3,
Grass 52.5,
Steel 51.4
(Note: Gyro Ball is counted as having 0 BP, which is what contributes to Steel being so Low. Grass knot is also counted as 0 BP)
Interesting questions.
Why are pokemon so weak to flying type moves and why are they so overused?
Does poison suck THAT badly?
Is the reason physical attacks used over special attacks because special attacks can be walled harder (I'm looking at you blissey)?
