Checks and Counters

Starting next month, I hope to add two sets of stats to my usage statistics thread: teammates (which Pokemon are most often paired with a given Pokemon on a team) and checks and counters.

Teammates is extremely straightforward (two Pokemon are either on the same team or they're not). But how do you define a check or a counter?

Well let's start by talking about the data we have access to. On its initial pass, my PO/PS log reader records information about the Pokemon that were present in the battle, but also the results of each and every matchup between two Pokemon.

So consider this example battle:

  1. Alice leads with Ferrothorn, Bob leads with Infernape.
  2. Alice pulls back Ferrothorn and sends out Heatran. Bob predicts that and uses Close Combat to KO.
  3. Alice sends out Gigalith. Bob uses Close Combat again, but Alice's Gigalith survives with Sturdy and goes for Explosion, taking out both Pokemon.
  4. Alice sends out Ferrothorn, Bob sends out Starmie.
  5. Alice switches Ferrothorn out for Lanturn while Bob switches Starmie out for Bouffalant.
  6. Alice has Lanturn use Volt Switch, swapping out Lanturn for Ferrothorn. Bouffalant attacks, not doing much to Ferrothorn.
  7. After incurring quite a bit of damage, Alice's Ferrothorn manages to KO Bouffalant.
  8. Bob sends out Scizor and takes out Ferrothorn with a U-Turn, sending Starmie out at the end of the turn.
  9. Alice's Starmie attacks Lanturn, activating Lanturn's Red Card and dragging out Scizor.
  10. Bob resigns.

My log reader would parse that battle into the following summary:

Ferrothorn vs. Infernape: Ferrothorn was switched out
Heatran vs. Infernape: Heatran was KOed
Gigalith vs. Infernape: double down
Ferrothorn vs. Starmie: double switch
Bouffalant vs. Lanturn: Lanturn was switched out
Bouffalant vs. Ferrothorn: Bouffalant was KOed
Ferrothorn vs. Scizor: Ferrothorn was u-turn KOed
Lanturn vs. Starmie: Starmie was forced out


Note that the ordering of the "x vs. y" is alphabetical, not by player. Also note the different types of ways a matchup can end:

  • double down: both Pokemon are fainted by the end of the turn
  • double switch: whether by u-turn or by actual switch, both Pokemon are replaced by the end of the turn
  • one Pokemon can be switched out
  • one Pokemon can be KOed
  • one Pokemon can KO the other while switching out (via U-Turn)
  • one Pokemon can be forced out by Red Card, Roar, etc.

So at the end of the month, after parsing hundreds of thousands of logs, we end up with a matrix of data telling us what happens whenever Pokemon x faces off against Pokemon y.

The question is, how do we use this data to determine who is the best counter or check for a given Pokemon?

But, BEFORE WE ANSWER THAT QUESTION, I have a much more fundamental question, which I want answered FIRST:

What is a check, and what is a counter? And how are they different?
 
Ok, I'll take a crack at the definition.
Counter: A Pokemon that beats all or most variations of another Pokemon. Capable of Switching in easily.

Check: A Pokemon that can impede another Pokemon's sweep or stall cycle, but may not always win if the opponent carries a certain move. This can also apply to moves or hazards.
 

Mr. Uncompetitive

That was when it all began.
is a Contributor Alumnus
From what I've learned so far in competitive battling (I'm not a newcomer but I'm not a veteran either, so I might be horribly wrong in some people's perspective), this what I define the two as:

A check is a Pokemon that is capable of being switched into a Pokemon in order to combat it, but it either fears some of the moves of the Pokemon it's switching into or it can't do much in return/has trouble beating it, therefore meaning the opponent MIGHT switch the Pokemon out. For example, Arceus-Steel is capable of checking Kyurem-W as it laughs at Draco Meteor and Ice Beam, as well as having the ability to 2HKO it, but it fears all of the coverage moves Kyurem runs: Fusion Flare, Earth Power, and/or Focus Blast. Another example is that CM Arceus-Steel is capable for checking Dialga as it doesn't care about anything other than Aura Sphere or Fire Blast outside of Rain, which aren't extremely common and it can be dealt with through CM and Recover. In return, however, Arceus-Steel can do little more than tell Dialga to get out since its two offensive moves, Judgement and Thunder, do next to nothing to Dialga unless it racks up a ton of CM boosts, burn with WoW and attempt to stall, or phaze with Roar. While Dialga can phaze it early on, it still can't do anything to Arceus-Steel without Aura Sphere or Fire Blast outside of Rain. and it doesn't stop Arceus from just coming back, meaning the Dialga user naturally will switch out to something more fit in dealing with Arceus-Steel, hence being a check.

A counter, on the other hand, is a Pokemon that doesn't really need to worry about the opposing Pokemon's most common moves and can actually beat the said Pokemon in return, meaning the opponent WILL switch out. So, for example, Arceus-Ground is capable of countering Thundurus-T since all it has to fear is the uncommon Grass Knot. In return, Arceus can lay waste to it with Ice Beam or Stone Edge.

EDIT: Yeah I guess Skarmory v. Metagross was a rather poor example (I don't even play OU regularly so what the hell do I know) so I'm changing it to another Ubers example to get my point across.
 

tehy

Banned deucer.
Skarmory is definitely a counter to metagross, mysteriousguy. Max defense skarmory takes 33% from thunderpunch. You say it can't do much but phaze it? True;but it can set up 3 spike layers while metagross is there, which means that the phazing will take off 25% of metagross' life, and it's so detrimental to the rest of the team that metagross will run. If metagross is running LO, then it does more, but then metagross is stallbait in any case. I don't even think CB 2hitKos. I'd argue that your counter definition totally fits skarmory anyways, since thunderpunch is pretty uncommon.

A check is something that CAN beat X pokemon. A counter is something that WILL beat X pokemon.
 
I'm barely beginning to play again, but I would have to define the following as:

Check: A pokemon that will come in not-so-easily, maybe after death fodder and threaten to out-beat the opposing pokemon whether immediatly or proceedingly. An example could be switching in a Swampert on a Charizard after it kills some other pokemon. Charizard can still do heavy damage, but you're opponent is swayed into switching because it will eventually outbeat it.

Counter: A pokemon that outright threatens the opposing pokemon with resistances, blocking strategies, set up windows and of course an easy KO.
An example of this could be a Blissey switching in on a DisableGar. Of course he can thunderbolt or focus blast all he wants but Blissey is using the opportunity to heal and set up. The more offensive side would be Choice Band Scizor switching in on a Mamoswine or Terrakion.

Both are meant to switch momentum in battle, and this is what I have understood from the two terms. Keep in mind you can use a counter to { v.check} another Pokemon:naughty:.
 
The range of a check and a counter is pretty wide in my opinion it can be very hard to pin down exactly what each is. Generally, a counter is a pokemon that can switch in once or more with little to no risk to itself and is capable of taking out the pokemon. A check is something that much switch in either on prediction of a move or after a pokemon is killed. This is a pretty skechy definition take for example the following situations:

1. Chansey into Yanmega: this is about as close as one can get to a true counter, Chansey littlerally fears nothing from yanmega, Chansey isn't worn down slowly, and yanmega can't do anything do it at all. Counters like this are rare.

2. Hippowodon into Lucario: Hippowdon can take anything (outside of high jump kick) Lucario has, and without spikes up, there is little chance of a 1HKO, and Hippowdon will just 1HKO Lucario if it attempts. Hippowdon can recovery any damage Lucario tries to put on it easily, meaning one on one, Lucario can't beat it. However, Lucario can bring Hippowdon to very low health with a +2 close combat, making it easy pickings for another sweeper. Hippowdon walls Lucario, but still has some risk. But because Hippowdon can beat Lucario, its still a solid counter.

3. Slowbro into Terrakion: Slowbro can switch into Terrakion and do massive damage to it. Against swords dance, Slowbro is 2HKOed, but still can pose a significant threat to Terrakion with scald or psychic / psycho shock. Its a similar story vs CB terrakion, slowbro can switch in, but still faces the risk of being slowly worn down and beaten. Slowbro can not always beat Terrakion one on one, but can still pose a threat to one or weaken it, generally Slowbro is considered a counter, a bit a shanky one, to Terrakion. (interestingly this draws parallels to the defensive check I mention later)

4. Chansey into Reunicles: basically, there are 2 types of reunicles, trick room, and calm mind. Chansey looses to calm mind, but not to trick room. It comes down to the set of the opposing pokemon here more than a specific move. Chansey is a solid counter for one, but loses to the other. To be honest, I don't known what you would call this.

5. Scizor vs Latios: Scizor can switch into Specs - Latios one time (unless its specially defensive with roost), on the second go it is finished off by pretty much whatever move Latios has. Scizor can in fact pursuit it to death, but if its lets say locked into surf, it can maybe 2HKO Scizor. Then it becomes a mind game, both pokemon can do significant damage to each other and it comes down to prediction who wins. (this is assuming no HP fire lol) The key here is Scizor can in fact switch in and maybe take it out, so this is again is usually considered a counter.

6. Focus sash Alakazam into Keldeo: because of focus sash here, Alakazam can come in pretty much all it wants and threaten to 1HKO with psychic, however, it can literally only do this once, that's it. Because of the fact that this is a one time guaranteed switch in, is it a check or counter? A fine line is probably drawn somewhere around here.

7. Mamoswine into Salamence: Salamence can do massive damage to Mamoswine if it switches in, but, Mamoswine if it predicts a dragon dance, and effectively come in, or after something has died, and threaten Salamence in return. This is probably as close to a true check as one can get.

8. Amoonguss into Thundurus-T: In this situation Amoonguss (if the spore if already fired off) will lose to Thundurus-T, but Amoonguss still takes a huge chunk out a Thundurus-S health, allowing something else to finish it off. This is what I consider a defensive check, other examples include defensive Politoed taking a hit a firing off a scald, or Forretress killing a dragon locking into outrage, they can finish or weaken a theat, but may lose to it.

What do you guys think? Where is the line? Pick apart my examples I am curious on how the community feels about each one.
 

ganj4lF

Nobody is safe from the power of science!
is a Team Rater Alumnus
Quoting the old CtP thread...

NixHex said:
To be considered a counter, a Pokemon must be able to switch into any of the opponent's moves under normal battle conditions*, including taking damage from Stealth Rock, and reliably KO the opponent before the counter is KOed itself. (props to Woodchuck)
-To be considered a check, a Pokemon must be able to switch into at least one but not all of the opponent's moves under normal battle conditions, including taking damage from Stealth Rock, and reliably KO the opponent before the check is KOed itself.

*Normal Battle Conditions consist of the following:
  • Stealth Rock is up on the counter's (your) side of the field.
  • Stealth Rock is not up on the featured Pokemon's (opponent's) side of the field.
  • The weather is handled as follows:
    • If no weather is listed, the weather is assumed to be clear skies unless the counter has the ability Sand Stream, Snow Warning, Drizzle, or Drought.
    • If the Pokemon to be countered can change the weather when a counter with a weather-changing ability comes in, the weather is situational and should be analyzed as such. For example, if the featured Pokemon is Drought Ninetales with Sunny Day and you list Politoed as your counter, you should evaluate the situation with two possiblities, rain or sun.
The definition of normal battle conditions is indeed arbitrary, but it seems reasonable to assume Stealth Rock is in play. For the weather issue, well it was explicitly stated in the CtP thread, but you'll need a different way to handle it here - for example by observing if weather is up or not in the actual battle. However I don't want to go too much into detail, since minor issues can be fixed later when you actually implement the algorithm; however I think that NixHex's definition is accurate enough to be utilized in such a setting.

However, there are some extreme cases in which the definition of counter can be less strict. Let me do a basic example: Offensive Heatran is indeed a counter to Specially Defensive Heatran, since it can smack it with Earth Power while taking literally nothing from the opponent. However, since SpD Heatran can take an Earth Power and Roar out the opposition, NixHex's definition is not clear on this case, since none of the two is actually KOed. So, a possible extension is that if a poke can deal heavy damage to the opposition without being crippled significantly in return, or being set-up on in some way, it is indeed a counter to that poke, even if it actually fails to KO it. FOr many (almost all) other cases, that definition should work. If there's some other case that's left out, we can discuss it here.

Hope it helped.
 
I think a counter to a pokemon not neccesairly means that it needs to kill the pokemon it´s supposed to counter but rather stop it in fullfilling it´s purpose in the team. For example deoxys-s´s purpose is to set up hazards/screens. So if you kill it but it still got SR/spikes up or dual sceens you didn´t countered it. But if you sent in your thundurus and prankster taunted it then you countered it. So my definition of counter an check would be:

Counter: a pokemon which under normal battle conditions can stop most variants of the pokemon from fullfilling its purpose inside the team.
Check: a pokemon which under normal battle conditions can stop some variants of the pokemon but will fail against others
 
6. Focus sash Alakazam into Keldeo: because of focus sash here, Alakazam can come in pretty much all it wants and threaten to 1HKO with psychic, however, it can literally only do this once, that's it. Because of the fact that this is a one time guaranteed switch in, is it a check or counter? A fine line is probably drawn somewhere around here.
Definitely a Revenge Kill.

Now that I look at what you guys are proposing, I think the definition I provided fit the key slot quite well.

I found this on neoseeker:

Examples of a counter: Chansey would be a 100% surefire counter to Unkown as even Choice Specs Hidden Power Fight fails to deal more than 15% damage while Chansey can Toxic stall or beat it with Seisemic Toss.

Rotom-w is a counter to Gliscor because of its ability to KO any Gliscor and Gliscor is unable to do anything to Rotom-w but use Facade for little damage.

A check is a Pokemon that can potentionally take out the said Pokemon it's supposed to get rid of, but is unable to assure success.

Examples of a check: Heatran is a check to Celebi as it can OHKO Celebi with STAB Fire Blast, however, it's not a counter because of the possibilty to Earth Power.

Infernape is a check to Tyranitar because despite resisting one of its STAB, Tyranitar can KO Infernape with Earthquake. Choice Band Stone Edge can KO Infernape, further reducing Infernape's ability to counter Tyranitar.

Another example of a check is Skarmory and Terrakion. Skarmory can switch in on any one of Terrakion's attack because of its near-impenetrable defenses, however, Terrakion can Swords Dance on the switch and deal heavy damage, if not KO, Skarmory with Life Orb boosted Close Combat.

Can we draw the line with the random definitions yet? They're more confusing than they are informative.
 
Examples of a check: Heatran is a check to Celebi as it can OHKO Celebi with STAB Fire Blast, however, it's not a counter because of the possibilty to Earth Power.
I don't feel that this is a check honestly. First off, if celebi is running earth power, Heatran can't ever come in and check it bar air balloon. So if a pokemon is running a specific uncommon coverage move for something specific, that pokemon no longer becomes a counter or even a check, just a dead pokemon.

I think the best way to put Heatran vs Celebi is a psuedeo or shanky counter. The act of switching in Heatran into Celebi, in theory fearing nothing if your risking it not having earth power, is infact countering Celebi. If Celebi has earthpower, it breaks the counter.

A beaten counter doesn't become a check. Heatran is a counter here, but Celebi can beat it.

Examples of this kinda beaten counter are signal beam Rotom-W to Celebi, heat wave bulk up Tornadus(-T) to skarm, and hidden power ice Venusaur to dragons or earthquake Venusaur to Heatran. None of these pokemon are checks at that point, but overall, they are solid counters outside of that 1/100 oddball.
 
I

A beaten counter doesn't become a check. Heatran is a counter here, but Celebi can beat it.
.
Beaten? The point is the intimidation you input with the switch in. Of course you're not an absolute counter to Celebi. Which is why you're a check. You're only beat when your Pokemon has fainted, and of course that is no longer a check so the proposal of beaten counters is quite absurd.

Hence the fact the term [check] comes from chess. You're putting the opponent in a situation where he can stay in for a risk or safely switch. An absolute counter is "switch out or die".

I don't see why this is so complicated.
Check: A pokemon that will come in not-so-easily, maybe after death fodder and threaten to out-beat the opposing pokemon whether immediatly or proceedingly.
Counter: A pokemon that outright threatens the opposing pokemon with resistances, blocking strategies, set up windows and of course an easy KO.
 

Jukain

!_!
is a Site Content Manager Alumnusis a Forum Moderator Alumnusis a Tiering Contributor Alumnusis a Top Contributor Alumnusis a Smogon Media Contributor Alumnus
okay so i've got some definitions for you:

counter: a pokemon that can handle every possible set of a pokemon, every possible variation, and has the ability to retaliate adequately

of course there are counters to single sets, but that is the best definition imho.

check: this is a bit broader, there are two types basically:
switch-in: a pokemon that can switch into one or more attacks of a pokemon, notably its primary stab, and outspeed or has priority to ko the pokemon from there or shut them down a la taunt thundurus vs hazards deoxys-s

revenge killer: a pokemon that can't necessarily switch in but can outspeed or has priority or can shut down a pokemon

these definitions account for bulky pokes which i see others don't.
 
Check: A pokemon that will come in not-so-easily, maybe after death fodder and threaten to out-beat the opposing pokemon whether immediatly or proceedingly.
Counter: A pokemon that outright threatens the opposing pokemon with resistances, blocking strategies, set up windows and of course an easy KO.
By your own definition of a check, Heatran is not a check. Heatran cannot come in on earth power Celebi, even if it does predict a leafstorm or an HP fire, or some in after a pokemon has been sacked. Celebi will just click earth power since it out speeds and proceed to smash through Heatran.

What you are proposing is almost 2 definitions of a check, the regular definition which you just posted, and pokemon which use the properties of a counter, which which you described, but fails at times.

It seems apparent to me that that is not the way of classifying Heatran vs Celebi. Heatran counters Celebi, if it fails, well then it fails end of story, it doesn't get demoted to a check.

What happens when I do the same to a check? The regular CB Scizor, with quick attack, checks a regular Volcarona after stealth rock damage for example. But if Volcarona is a bulky version with morning sun, it can take a quick attack from Scizor and heal up health or even just spam morning sun until Scizor gets burned. What is CB Scizor then, it has failed to adequately check an uncommon Volcarona, does it drop to some sort of lower check status as a whole then? No, it’s a regular check to Volc with flaws; Heatran is a regular counter to Celebi with flaws.
 
Did anybody notice I found that example on neoseeker?

If I'm not defending it adequately it's because it wasn't my proposal to begin with. Nevertheless, a move-scouted Celebi lacking Earth Power can be checked by a Heatran. It's not a matter of demoting, but Heatran simply isn't an outright counter for Celebi, since it does not make it pale of the sight of every single one or even a majority of its sets. SO, I'd have to agree.

Revenge killers:
Dugtrio
DestinyGar
Alakazam
Checks:
Mamoswine v Salamence
Electivire v Gyarados
Counter:
DisableGar v Medicham
Tangrowth v Swampert
Blissey v Celebi
 
A true counter can switch into the threat and consistently come out on top. 1 vs 1, they win the fight.

A check is a bit looser. It can come in and threaten the threat (be it by tanking a hit on the switch and outspeeding ftw, or tanking 2 hits and KOing, or tanking X hits and YKOing. A revenge killer can also "check" something.) The difference here is that the check will win 1 vs 1, but either A) can't necessarily switch in and win, or B) will be left severely weakened by doing so.

Both checks and counters force the threat out. However, a counter will be able to do so consistently throughout the battle, while a check will come out on top, but can't be counted on for the long haul. A threat can slowly wear down and break it's own checks. It ain't getting past a counter.
 
From what I'm reading, it sounds like this post sums up the definition and difference quite nicely.

So now we move onto the theorymonning section. Given the matchup:

Pokemon A vs. Pokemon B

There are nine different possible outcomes:

  1. Pokemon A is KOed
  2. Pokemon B is KOed
  3. Both Pokemon are KOed
  4. Pokemon A is switched out (or u-turns out)
  5. Pokemon B is switched out (or u-turns out)
  6. Both Pokemon are switched out
  7. Pokemon A is forced out*
  8. Pokemon B is forced out*
  9. Pokemon A u-turn KOs Pokemon B**
  10. Pokemon B u-turn KOs Pokemon A**

*I don't think it's possible for both Pokemon to be forced out on the same turn.
  • Terrakion vs. Metagross: Terrakion switches out into Forretress
  • Metagross attacks and dies from Life Orb recoil
  • Metagross is replaced by Magnezone
This counts as a u-turn KO for Terrakion.

You probably disagree with this assessment, but what you're actually disagreeing about is that I call it a "u-turn KO," rather than something like a "switch KO."


So, if we make pie charts of the frequency of occurrence of each category, how do we expect the pie charts to differ if Pokemon B is a counter for Pokemon A, a check for Pokemon A or neither?

Keep in mind that I do not yet have the ability to do "1337" stats (damn Zarel and his ladder changes...), so we expect a good amount of "nosie" due to unskilled players.
 

ganj4lF

Nobody is safe from the power of science!
is a Team Rater Alumnus
If B is a counter for A, we could probably expect possibilities 1, 4 and 7 to be significantly higher than "normal". Also, 2 and 9 should not happen at all, since if we're considering the definition of counter I gave earlier, they are contradictory with that so given that B is a counter for A, they shouldn't happen (or should happen very rarely, since critical hits, parahax, misses and such things happen). Okay, this is probably an oversemplification since if the counter is weakened enough when it comes in, option 2 becomes possible, but I think you got the point anyway - 2 and 9 should be lower. 3 is possible, but should be quite rare too, for the same reason 2 should be rare. 8 shouldn't happen at all. 10 is possible, and can be higher than normal, it basically is a sub-point of 1. Points 5 and 6 are a bit tricky, since prediction comes into play, and double switches are indeed possible. I really don't know what to expect about those points.

If B is only a check, we can infer much less information, since it is indeed possible for B to just switch into the wrong move and be KO'd. Imagine Jellicent switching into a CB Stone Edge from Terrakion while it expects a Close Combat - it can be easily 2HKO'd while still being a check to Terrakion, and quite a good one. Anyway, the fact that B is a check should increase somewhat 1, 4 and 7 as in the precedent case. It won't lower too much 2, 5 and 8, especially if the level of the players is high, since prediction will be more frequent and effective. 3 and 6 are both tricky, since it boils down to what specific check we're considering - if you're using your own Heatran to counter opponent's SpD Heatran, 6 is very likely to happen; 3 is not so frequent, so I cannot way much on that. 9 and 10 are sub-cases of 1 and 2, so they should be treated as such.
 

bugmaniacbob

Was fun while it lasted
is an Artist Alumnusis a CAP Contributor Alumnusis a Top Contributor Alumnusis a Top Smogon Media Contributor Alumnus
I don't have much to add myself but CAP's definition of a Counter might be worth a look for your purposes

Essentially there are six characteristics of a counter, as follows:
  • Can switch into the Pokemon's strongest reliable STAB attacks at least three times from full health.
  • Can switch into the Pokemon's strongest possible coverage move at least twice from full health.
  • Can stall the Pokemon indefinitely using its recovery options, either forcing the Pokemon out or healing enough that the Counter can alternate between recovery and attacking.
  • Can OHKO or 2HKO the Pokemon with one of the moves on the Counter's relevant official Smogon moveset.
  • Can cripple the Pokemon with a permanent status move without risking an OHKO.
  • Can set up, use hazards, weather, or otherwise execute an opponent's strategy without risking a 2HKO.
A Pokemon that fulfils three or more of these criteria is a hard counter, and one or two of these, a check.

Dunno how much that helps but it's suited the CAP Project's purposes thus far. It doesn't mention common battle conditions, but I'm sure you could probably add "under common battle conditions" to all of the above if it bothers you that much.
 

Stallion

Tree Young
is a Tiering Contributoris a Battle Simulator Moderator Alumnusis a Three-Time Past WCoP Champion
A check is something that can't switch in (or if it can, it can only be on a particular move), but it prevents a threat from sweeping.

Example: Scizor is not a Terrakion counter, because it gets raped by both STAB moves. Barring the rare Babiri berry, Scizor's Bullet Punch is used to guarantee that Terrakion won't sweep you though and is known as a check.

Example 2: Scarf Jirachi isn't a counter to Dragonite, as Choice Band Earthquake or Fire Punch do a number to it. It can take at least one hit from any move though, and outspeeds it after a Dragon Dance and hits it with Ice Punch/Iron Head, so it is a check.

Example 3: Sash Alakazam can be seen as a check to Lucario, as it can come in after a teammate has fallen, take an Extremespeed and KO it. It cannot be seen as a counter though, as literally any move followed by an Extremespeed means that Lucario comes out on top.

People saying that Heatran is a check to Celebi is stupid unless you're talking about Scarfed Heatran, it's either a guaranteed counter (Earth Power-less versions) or it is not even a check, because no matter what it switches into it'll die to Earth Power the next turn. A non scarfed Heatran should be seen as a weak/shaky counter, as should similar examples of that nature.
 
Okay guys, I think we're good in terms of definitions. Anyone have any theories on how a check will differ from a counter in terms of the measurable quantities outlined above?

In a few days, I hope to be able to test some of your theories, but for now, there a few glitches I need to work out with my log parser.

Namely, this.
 

eric the espeon

maybe I just misunderstood
is a Forum Moderator Alumnusis a Researcher Alumnusis a Top CAP Contributor Alumnusis a Tiering Contributor Alumnusis a Top Contributor Alumnus
Something which I've always found key to the difference between a check and a counter is whether the Poke can only come in after a KO/u-turn. Would it be practical to check how the checking/countering Pokémon switched in as well as how that matchup ended?

Anyway, onto what you're requesting. I've grouped the ones where each has the advantage since the only sane way to tell what proportion of the matchups where a Pokémon is at a disadvantage leads to it being switched out and what proportion are left in to faint as a sacrifice is by looking at the stats you'll generate.

A counters B

Pokemon A is KOed/switched out (or u-turns out)/u-turn KOed by Pokemon B - <10%
Pokemon B is KOed/switched out (or u-turns out)/u-turn KOed by Pokemon A - >85%
Both Pokemon are KOed - <15%

Forcing out is questionable for counters, you could argue that just Roar/WWing out whatever is being faced (either way) does not really counter something/prevent something being a counter. It'd be too easy for something like Skarmory to apparently have no counters because it forces things out often, even though in reality a large number of Pokemon can do what most players would consider countering (switch in safely, threaten to OHKO/2HKO). I would not count these outcomes at all.

Allowing 15% loss for the countering 'mon is a guess at how often the counter will be sacrificed/the 'mon to be countered will already have set up against something else to the point that it is uncounterable, when you run the numbers this may of course turn out to be wide of the mark.

A checks B

Pokemon A is KOed/switched out (or u-turns out)/u-turn KOed by Pokemon B - <25%
Pokemon B is KOed/switched out (or u-turns out)/u-turn KOed by Pokemon A/is forced out/Both Pokemon are KOed - >75%


Forcing out is a fine way of checking something, being forced out does not really stop you being a check. Both Pokemon being KOed indicates to me that a check has done its job, for example by Exploding on the target and killing it.

Both Pokemon are switched out - Probably should be discounted for all, since a large proportion will be due to the prediction of a switch by the opponent. Does not give much information as to which active poke has the advantage, unless you can tell which Pokémon switched in most recently.

Additionally, I'd suggest a category for trappers which distinguishes between Pokes that they kill and those they just have the advantage against but can switch. Something like:

A trap-kills B

Pokemon A is KOed/switched out (or u-turns out)/u-turn KOed by Pokemon B/Pokemon B is switched out (or u-turns out) - <15%
Pokemon B is KOed/u-turn KOed by Pokemon A - >80%
Both Pokemon are KOed - <20%

Slight leniency in letting Pokémon escape because Shed Shell. And maybe trap-kills is not quite the right name, but it would be interesting to see which Pokémon actually get the kill against which as well as which force eachother out.

Extra information which could be very helpful:
The health (HP and maybe status) of both Pokémon at the start of the matchup (maybe discount pokes below a certain HP which switch in and die without moving, since they're sacrifices). With this adjustment I'd want to change my numbers, make them much more strict.
How the matchup started (which poke(s) was switched in, and whether they had to risk an attack from the other doing so).

*I don't think it's possible for both Pokemon to be forced out on the same turn.
Would Dragon Tail vs a Red Card holder not do this?
 
Okeydoke. I've sorted out my log-parsing issues (by defining "fodder," which consists of Pokemon A switching into Pokemon B and dying from entry hazards or other passive damage).

So now I've got an "encounter matrix" I can use to provide you guys with some raw data. My sample dataset is OU battles from 10/1 - 10/18.

I'm going to start with Scarfwynaut's list of checks and counters. I'm also just gonna post the results. Gonna leave it to you guys to decide what the data means.

Also, could someone with a background in statistics could tell me how to calculate "margin of error" for this kind of data?

1. Chansey into Yanmega: this is about as close as one can get to a true counter, Chansey littlerally fears nothing from yanmega, Chansey isn't worn down slowly, and yanmega can't do anything do it at all. Counters like this are rare.
  • Yanmega was KOed 11 times
  • Blissey switched out 10 times
  • Yanmega switched out 15 times
  • Double switch once

2. Hippowodon into Lucario: Hippowdon can take anything (outside of high jump kick) Lucario has, and without spikes up, there is little chance of a 1HKO, and Hippowdon will just 1HKO Lucario if it attempts. Hippowdon can recovery any damage Lucario tries to put on it easily, meaning one on one, Lucario can't beat it. However, Lucario can bring Hippowdon to very low health with a +2 close combat, making it easy pickings for another sweeper. Hippowdon walls Lucario, but still has some risk. But because Hippowdon can beat Lucario, its still a solid counter.
  • Hippowdon was KOed 88 times
  • Lucario was KOed 114 times
  • Double down three times
  • Hippowdon switched out 36 times
  • Lucario switched out 93 times
  • Double switch 4 times
  • Hippowdon was forced out once
  • Lucario was forced out 25 times
  • Hippowdon was U-Turn KOed 3 times*

*How can Lucario U-Turn KO anyone? That's because roar-then-dies-by-residual-damage count as a U-Turn KOs for the opponent. Again, this is part of the definition. If you don't like it, call it something else in your head.

3. Slowbro into Terrakion: Slowbro can switch into Terrakion and do massive damage to it. Against swords dance, Slowbro is 2HKOed, but still can pose a significant threat to Terrakion with scald or psychic / psycho shock. Its a similar story vs CB terrakion, slowbro can switch in, but still faces the risk of being slowly worn down and beaten. Slowbro can not always beat Terrakion one on one, but can still pose a threat to one or weaken it, generally Slowbro is considered a counter, a bit a shanky one, to Terrakion. (interestingly this draws parallels to the defensive check I mention later)
  • Slowbro is KOed 54 times
  • Terrakion is KOed 88 times
  • Double down twice
  • Slowbro switched out 74 times
  • Terrakion switched out 162 times
  • Double switch 14 times
  • Slowbro was forced out once
  • Terrakion was U-Turn KOed 3 times

4. Chansey into Reunicles: basically, there are 2 types of reunicles, trick room, and calm mind. Chansey looses to calm mind, but not to trick room. It comes down to the set of the opposing pokemon here more than a specific move. Chansey is a solid counter for one, but loses to the other. To be honest, I don't known what you would call this.
  • Chansey was KOed 49 times
  • Reuniclus was KOed 29 times
  • Chansey switched out 170 times
  • Reuniclus switched out 109 times
  • Double switch 18 times

5. Scizor vs Latios: Scizor can switch into Specs - Latios one time (unless its specially defensive with roost), on the second go it is finished off by pretty much whatever move Latios has. Scizor can in fact pursuit it to death, but if its lets say locked into surf, it can maybe 2HKO Scizor. Then it becomes a mind game, both pokemon can do significant damage to each other and it comes down to prediction who wins. (this is assuming no HP fire lol) The key here is Scizor can in fact switch in and maybe take it out, so this is again is usually considered a counter.
  • Scizor was KOed 680 times
  • Latios was KOed 859 times
  • Double down 27 times
  • Scizor switched out 215 times
  • Latios switched out 390 times
  • Double switch 181 times
  • Scizor was U-turn KOed 2 times
  • Latios was U-turn KOed 158 times

6. Focus sash Alakazam into Keldeo: because of focus sash here, Alakazam can come in pretty much all it wants and threaten to 1HKO with psychic, however, it can literally only do this once, that's it. Because of the fact that this is a one time guaranteed switch in, is it a check or counter? A fine line is probably drawn somewhere around here.
Keep in mind that I can't (yet?) break this stuff down by movesets, so these stats are for ALL Alakazam, and only ~40% of Alakazams are sashed.

  • Alakazam was KOed 119 times
  • Keldeo was KOed 299 times
  • Double down 4 times
  • Alakazam switched out 10 times
  • Keldeo switched out 250 times
  • Double switch once
  • Alakazam was U-turn KOed once

7. Mamoswine into Salamence: Salamence can do massive damage to Mamoswine if it switches in, but, Mamoswine if it predicts a dragon dance, and effectively come in, or after something has died, and threaten Salamence in return. This is probably as close to a true check as one can get.
  • Mamoswine was KOed 93 times
  • Salamence was KOed 676 times
  • Double down 7 times
  • Mamoswine switched out 43 times
  • Salamence switched out 384 times
  • Double switch 13 times
  • Mamoswine was forced out 3 times
  • Mamoswine was U-Turn KOed 3 times

8. Amoonguss into Thundurus-T: In this situation Amoonguss (if the spore if already fired off) will lose to Thundurus-T, but Amoonguss still takes a huge chunk out a Thundurus-S health, allowing something else to finish it off. This is what I consider a defensive check, other examples include defensive Politoed taking a hit a firing off a scald, or Forretress killing a dragon locking into outrage, they can finish or weaken a theat, but may lose to it.
  • Amoonguss was KOed 21 times
  • Thundurus-T was KOed 21 times
  • Amoonguss switched out 45 times
  • Thundurus-T switched out 63 times
  • Double switch 19 times
  • Amoonguss was U-Turn KOed 1 time
  • Thundurus-T was U-Turn KOed 3 times
 
Also, could someone with a background in statistics could tell me how to calculate "margin of error" for this kind of data?
You could view these battles with one specific match-up as a set of tests you are running, and specific outcomes can be chosen as either "success" or "failure" (at countering the pokemon, that is). If you do this, it would be quite easy to give you a decent estimate of the success-rate + a margin of error (depending to the level you significance you choose)with a "test on a percentage"; However, this isn't perfect of course since forcing something out is "more success" than outright killing it, while stuff like double switches /double kills have to be neglected. But it'll be most likely be usable.

The other thing you could do would assign numbers to every outcome( for example 6 for "killing" , 4 for "opponent forced out" 3 for "opponent switched out" and vice versa -6 for " killed" usw). Then, you could run a "t-test" on all battles with a specific match-up and it will give you estimate of the average outcome + a margin of error(still depending to the level you significance you choose). Than, you have to decide how to use the number you got (for example in clinical studies, you would choose a critical point, and then look whether your average outcome minus the margin is higher than it or not; If yes, it's considered "good enough", if not, not)
This, on the other hand, can be very biased depending on how you assign the numbers and how you choose to process these numbers, so I wouldn't propose this way - you could end up with everything being a counter to everything, stuff like that. Though, if you are determined at spending much time with it and trying out different ways of assigning & processing, it COULD get better than the other test.

On a sidenote:
On lesser used pokemon, the margin of error will most probably be so big that everything you do is most likely rubbish. As a rule of thumb, you should use these statistics only if n*p*(1-p)>9 applies, with "n" being the number of battles against one specific pokemon and "p" being the estimated success rate. So, if you estimate the success-rate at 75%, n should be greater than around 50 and so on.

I could either try to describe these specific test statistics to you or do it myself(if I find the time, though), but describing is somewhat difficult because I'm no native speaker and writing about mathematics is especially difficult in a language you didn't learn it in since you don't know which words are commonly used in respective professional circles.
 
haxcho, I'm asking for something a bit simpler:

n measurements, m possible outcomes: given frequency counts f_1-f_m, where sum{i}f_i=n, what is the margin of error on each f_i?
 

Users Who Are Viewing This Thread (Users: 1, Guests: 0)

Top