Articles

Published on June 17th, 2014 | by bearsfan092

14

2014 May International Challenge Retrospective

The dust has settled on the first Wi-Fi tournament with something at stake and sign-ups for the next are already open. After the debacle that was last year’s Wi-Fi system, it was pretty clear that adjustments needed to be made in order to continue distributing Championship Points using Wi-Fi tournaments. I’m pleased to say that the adjustments were made, and for the most part, results were positive! Of course, a new system brings new lessons, and I’ll examine some of the new aspects of this year’s Wi-Fi tournament structure.

The Battle Limit

For me personally, this is the greatest change. Previous tournament winners were more or less determined by who could play the most games with a certain amount of consistency. With the battle cap, this is taken out of the equation. This encourages more players to partake in a competition that otherwise would not, for fear of not being able to play enough games.  To illustrate my point, I’ll cite some figures provided by AlphaZealot from last June’s International Challenge:

Ranking Average Games Played
Top 5 146.2
Top 10 113.8
Top 25 100.68
Top 50 84.8
Top 75 79.44
Top 100 75.77
Top 250 61.73
Top 500 54.158

As you can see, the average top five player battled in almost three times as many games as players in the top 500 group.  More practice does generally result in a better player, but it’s difficult to say a player is so much better than another when playing so many games is necessary to achieve a good placing.  At that point, winners seem to be determined by who sets apart more time for Pokémon more than actual skill. (Please note: This is not to detract from our top five players’ successes by any means).  For that reason, I really like what TPCi is attempting to do with the battle limit.

However I would be remiss if I did not discuss the drawbacks of this system. With 60 battles, randomness (and not just the “hax” variety) becomes more noticeable. Of course, if you get a bad stroke of luck, you just used up one of your precious 60 games on a loss you might usually win. However, we also need to consider the opponent matchmaking system, which I will expand on in another section. If you get paired up with someone >50 points below you, you have less to gain and more to lose. If you win, you use up one of your 60 games for a 5-10 point boost. If you lose, you just lost 20 points and one of your games. Of course if you just get matched up with players who are roughly around your rating, then you’ll probably just lose more anyway, so the matchmaking system, at least probably, won’t skew results too far from what they would have been otherwise. Overall, I think AlphaZealot and TPCi should take pride in introducing the battle limit. Perhaps some fine tuning may need to be made, but the system largely works.

Kangaskhan / Smeargle: Autolosses in Wi-Fi

Once again, I’m going to stress that there are only 60 games to achieve a rating with.  Now let’s suppose you managed to build a team that takes on the entirety of the metagame with a reasonable matchup save for one combination – Kangaskhan and Smeargle.  I’m going to take it a step further and assume that your team automatically loses to this combination, just for the sake of debate.  This duo can probably trigger an automatic loss faster than anything else in the metagame if the opposing team is not prepared in advance, which is why I use it as an example.  If you were unfortunate enough to have such a bad matchup against this combo, expect your Wi-Fi performance to take a serious hit.

It’s not so much a statement about Kangaskhan-Smeargle in particular as it is about automatic losses. I will beat this into your head.  You have only 60 games to finish in good standing.  If you go into team preview and see a combination that you cannot win against, you just burned up one of your precious games and suffered a rating hit.  Worse yet if it’s a remotely common matchup like Kangaskhan-Smeargle.  In that case, you can reasonably expect to see any common combination 2-3 times in a single tournament. For the sake of argument, let’s assume you lose an average of 20 points for every loss.  This means that you’re giving up 40-60 points for free in a given tournament.

I think that this feature of this year’s Wi-Fi tournaments separates it from last year’s.  I’m going to arbitrarily and hypothetically say that a 90 percent win-loss ratio is necessary to reach into the top 4.  In 2013, you could (again, hypothetically) build a team that matches up well with 90 percent of the metagame, but absolutely cannot win against the remaining ten percent.  After all, if you ran into one of your autolosses, you could just get back on the ladder and most likely win enough matches to cover up the team shortcoming.  If you were unlucky enough to hit the same bad matchup within a short time span, it still wasn’t a big deal if you could play enough matches to cover both of these losses.

In 2014, we don’t have that luxury, or at least not to the same extent.  If we use the same hypothetical ratio as the 2013 example, you need to win 9 games to cover up an autoloss.  The difficulty appears if you happen to get the same bad matchup in a short duration.  For instance, say you lose game one on a bad matchup.  You follow up with wins in games 2 through 5, but you grind your teeth as your autoloss stares you in the face during game 6’s team preview.  Now you have two losses due to teambuilding, and now you need to win the rest of your 20 games that day to maintain the 90 percent ratio.  Assuming that future tournaments use the 20 battle per day limit, that’s a whole day spent recovering from one matchup you overlooked!

Personally speaking, I think this is one shortcoming I had in the May International Challenge.  In my particular case, the Kangaskhan-Smeargle combination was not an autoloss strictly speaking, but it was a matchup that I had to steer very carefully at the beginning with a very specific lead combination or else I would lose the game by turn 3.  Overall, I went 2-2 against Kangaskhan-Smeargle.  Considering that one of my losses was to a player over 100 points below me, I believe that these two losses had a significant impact on my rating.  Had I prepared better, perhaps I could have finished higher. Maybe in your case, you might find that you can’t beat Amoonguss alongside a setup Pokemon like Tyranitar, Kangaskhan, or Azumarill.  Whatever your autoloss is, I highly suggest addressing it, because it may end up costing your rating dearly.

Rating Matchups

I think one of the more intimidating things for someone at a Pokémon event with Swiss is the prospect of being paired down. Since it’s very easy to lose a single game of Pokémon, the penalties for losing are much scarier. Fortunately, pair downs at live Swiss events only occur when all other options have been exhausted. This means the relative frequency of pair downs make them a necessary but tolerable evil. Strictly speaking, pair downs don’t happen in Wi-Fi tournaments where an Elo system is used rather than a Swiss record. However, the situation is similar. Losing a matchup causes your rating to tank. Winning results in a lesser reward. In a Swiss system, a win in a pair down still holds significant reward as it goes toward record, which is the primary seeding mechanism in Swiss. The problem with Wi-Fi tournaments is that the frequency of pair downs is really, really high. In addition, the magnitude of the pair down can be pretty large. I’ve said before that if you lose, you take a large penalty. If you win, you don’t gain as much. With the inherent volatility of Pokémon, it’s very easy to drop a game or two to someone of much lower rating over the course of 60 games. On top of that, you’re using your 60 battles to beat up on lesser rated players. If someone else’s luck of the draw yields higher rated opponents, they may finish with a significantly higher rating. The solution to this is to tune up the pairing criteria of the system. There should be no reason that players with a rating disparity greater than 100 points should be playing each other. Putting in a rating difference limit would reduce the frequency of these matchup. This has the additional benefit of pairing up players of similar skill level, which makes everyone happier. Of course, a rating difference limit falls apart at the extreme ends of the rating spectrum, but these extremes can be accounted for in the formula. Really, I just don’t want to see players sitting around 1750 flipping tables because they had a streak of unlucky rolls against a 1600 player, and I don’t think anyone wants to see that either.

Time Limits

This isn’t really new nor is it exclusive to Wi-Fi tournaments, but it does deserve a brief mention. 15 minute game timer with a 45 second turn timer can result in a timeout.  However, this isn’t to the same extent as 2012, so if my experiences of this tournament are indicative, the timer will do its job in discouraging stall strategies such as Minimize and Stockpile. It also allows for two players to run fairly defensive teams and finish most matches in time. I think moving forward, players should be mindful of how long it takes for them to win. Any win condition that takes longer than a Leech Seed checkmate should be considered too slow, and even Leech Seed needs to be executed within a certain time limit. In short, play to knock out your opponent and you should be fine.

The Disconnection System

I started this article on a positive point, and I’m going to send you off on a positive point. I’m going to be dramatic and say that disconnections were a huge blight on the legitimacy of last year’s Wi-Fi tournaments.  For those not familiar, the old system could not discern who disconnected, so wins nor losses were awarded. Instead, the disconnection rate of both parties was increased, and those above a certain threshold were deemed frequent disconnectors and disqualified. Naturally, there was a possibility of taking out innocent parties, and I’m inclined to believe that this happened at least once. The improved technology behind the Wi-Fi system of the sixth generation has largely eliminated this problem. Connection stability seems greatly improved over last generation, albeit not perfect for everyone. Disconnections usually result in appropriate rewards and penalties for both parties. This makes the results much more reflective of what actually happened during the tournament, and that can only be a good thing. I take my hat off to everyone who made the new system possible. Of course, there are some minor issues to be handled. Rumors, and I do emphasize that they are rumors, of disconnects without reward or penalty have been circulating. Similar stories have proven true for past Nintendo Wi-Fi games such as Mario Kart DS and Metroid Prime: Hunters, both of which penalized disconnections. However, do note that all stories are so far hearsay. Of course, I encourage anyone who had this happen to them to speak up, but only if proof can be shown. I’d hate to have a good old fifth generation disconnection witch hunt.

Closing out, I think I’m justified in saying that May International Challenge was a success, and it convinced me that Wi-Fi tournaments can work for CP distribution.  In the meantime, feel free to post your thoughts in the discussion, and we’ll see you for the next International Challenge!


About the Author

is an avid lover of Jigglypuff and all things VGC. He also runs a youtube channel named "bearsfan092drums", which hosts a series of RNG abuse tutorials. He recently won the 2012 Philadelphia Fall Regional for VGC.



14 Responses to 2014 May International Challenge Retrospective

  1. melevin9 says:

    Good shout on kangaskan-smeargle considering that 4 kangaskhan-smeargle teams placed inside the overall top 10 in the May IC

  2. Scott says:

    Really, I just don’t want to see players sitting around 1750 flipping tables because they had a streak of unlucky rolls against a 1600 player, and I don’t think anyone wants to see that either.

     
    I think this is just the game though — I don’t know if you quite hit the metaphor in the article, but I think the impact on your overall placement is more similar to losing a game against someone who has a much better record than they should have during Swiss or whatever. The punishment is similar — they are almost definitely going to lose more games than they win in the remaining matches, so the loss effectively counts for about double what it normally would because it tanks your opponents’ win percentage. The point of me writing this being that like most of the other complaints, something analogous to what happens on Wi-Fi can happen in real life too. A lot of Pokemon, for better or worse, is not losing the bad ones, regardless of the format of the event. I think that’s something I wish we’d have had more Wi-Fi tournaments this year to help people learn — I know I learned an awful lot about how not to lose in round 1-4 of real life events last year from how I had to play in Wi-Fi tournaments.
     
     
     
    I think with that in mind, the idea of ignoring specific matchups like the Kangaskhan/Smeargle matchup is interesting. I think in preparation for Nationals and Worlds, most of the people who do consistently well and who make their team choices more on being anti-metagame rather than on just doing their own thing tend to ignore matchups they deem improbable or unimportant pretty regularly. I think this is pretty obviously a strategy that works for the big live events — Ray preaches this line of thought pretty hard, for instance — but I think it is a much weaker option strategically in a Wi-Fi tournament. There’s a big difference in what you’re going to see if you’re pulling the random opponent slot machine arm down 60 times instead of 8, and even with Wi-Fi tournaments being pretty abbreviated compared to last year, you’re almost definitely going to hit anything that’s even remotely popular. There’s sort of a line there — what is too obscure to be ignored for a tournament like this and what isn’t? Obviously based on the previous tournament, Smeargle is not so obscure that it can be ignored, and I think if you want a bunch of points for the next one, you’ve gotta have a good plan for it. I think it’s still important to figure out what you can ignore though, and what you can afford to take weaker matchups against. Perish Trap? Harder Trick Room? Rain? Sun? Smeargle? You know you’re going to face an abundance of goodstuffs offensive crap around Kangaskhan/Talonflame and apparently Smeargle, you know you’re gonna see a lot of suddenly threatening badstuffs offense with Greninja/Talonflame/(mega), what aren’t you going to see much of? How many times are you willing to play a bad matchup to sacrifice it to improve your matchups against more common archetypes?
     
    I think it’s an interesting series of questions. I think it’s a little interesting especially in reference to you asking about using real teams or not for the next Wi-Fi event on Twitter, because I don’t think the way I’d play for matchups if I wanted to win a Wi-Fi tournament is quite the same as what I’d play if I wanted to win Nationals anyway. I’d be pick-and-choosing what I wanted to counter a lot more carefully at the National level. I don’t think Wi-Fi tournaments are something I’d just throw a team I’d want at a live event at unless I was doing it for practice to make myself more ready for the real events rather than to optimize Wi-Fi points, since you both need to win faster to avoid the RNG over the big sample of games and have better overall matchups to avoid bad matchups to optimize points on Wi-Fi. That said, I’d advise most people to just use their Nationals team to be 60 games of higher pressure practice better than their peers who are denying themselves practice to dodge probably irrelevant scouting anyway, since I don’t think it’s likely this Wi-Fi tournament will matter for 99.9% of players, but still.

  3. NinjaSyao says:

    Regarding the rating matchup I wholeheartedly agree. Got paired up a LOT with players either 100 points or more below or above me and this sucked with needing to win 3-4 times to make up for 1 loss each time. A closer pairing criteria would definetly help with this and make it a more even playing field for players and thus more fun as you said.
     
    Autolosses… the IC really puts your team through the wringer with the amount of newcomers entering and thus a great amount of gimmicks/singles tactics/crap being present, but it does really help in identifying your autoloss combo and how well your team deals with onorthodox stuff. Despite me thinking otherwise with 2 priority users on the team it turned out that Khan and Smeargle was my autoloss. Got mauled a couple of times and still had to play really well if only one of them was present on the team. Quit at the end of Day 2 after realising this flaw in my team, but now got a counter on the team so feeling ready for this IC! :D
     
    Was kinda hoping they would do one of these each month as they´re great practice for helping you grow as a player. For the rest this new way of handling the IC is definetly an improvement.

  4. Queejibo says:

    I think the reason some people think that a disconnection sometimes doesn’t affect the score system is that it often doesn’t get updated immediately. A lot of the times when I play against somebody on the rated battle spot and he/she disconnects, the score will appear the same. It will fix itself after another battle or something.

  5. R Inanimate says:

    Some good thoughts here, and some pretty decent advice on not having hard autolosses just as we are about to get the June IC underway.
     
    Regarding matchmaking… from what I’ve heard, it tries to pair you up with someone relatively close to your ranking first, but if nothing bites the fishing line after a while, it will broaden the matchmaking range (and will cancel out if it still can’t find anything). Because of this, it can be a bad thing to try to get too far ahead of the pack too quickly, as you may end up constantly being unable to find people with ratings closer to where you are. While things will always depend on who’s online and where they are at, it seems like the tourney either encourages people to pace themselve and spread out their battless a bit, or to go on a 60 battle marathon on the last couple of days if they want a better shot at a good rating.
     
    I also kind of doubt a 90% Win Rate is all that necessary for T4. I think I was 4th last time by a few points and my W/L was closer to 80% than 90%.

  6. “I’m going to arbitrarily and hypothetically say that a 90 percent win-loss ratio is necessary to reach into the top 4.”

    I collected a bit of data regarding this after the tournament.

    4th – R Inanimate – 1874 (41-12) [77.4%]
    9th – joej m – 1865 (46-14) [76.7%]
    20th – Birch – 1832 (35-10) [77.8%]
    26th – Primitive – 1827 (48-12) [80.0%]
    31st – bearsfan – 1821 (35-9) [79.6%]
    50th – LPFan – 1809 (36-10) [78.3%]
    65th – Mrbopper – 1804 (40-16) [71.4%]
    69th – kingofmars – 1802 (43-16-1) [72.9% excluding the tie]

    A interesting point regarding the Autolosses section of your article is you actually ended up with a very high win percentage. Although at the same time it can be seen that your score seemed to suffer from what I would assume are big hits in your losses.

    One topic that wasn’t address in the article that I haven’t heard much people talk about is the CP payout. Right now Wi-fi is worth less than a Premier Challenge. The reasons for the low payout was obvious last year due to legitimacy issue. However, with some of the larger options fixed, should they now be worth more and how much?

  7. bearsfan092 says:

    I know I learned an awful lot about how not to lose in round 1-4 of real life events last year from how I had to play in Wi-Fi tournaments.
     
    This sentence really stuck with me in this paragraph, and it’s one of the reasons I’m going to play all 60 matches this tournament regardless of rating.  These tournaments teach you so much about how to approach unknown situations under pressure (although not as high as a live event).
     
    I think with that in mind, the idea of ignoring specific matchups like the Kangaskhan/Smeargle matchup is interesting. I think in preparation for Nationals and Worlds, most of the people who do consistently well and who make their team choices more on being anti-metagame rather than on just doing their own thing tend to ignore matchups they deem improbable or unimportant pretty regularly. I think this is pretty obviously a strategy that works for the big live events — Ray preaches this line of thought pretty hard, for instance — but I think it is a much weaker option strategically in a Wi-Fi tournament. There’s a big difference in what you’re going to see if you’re pulling the random opponent slot machine arm down 60 times instead of 8, and even with Wi-Fi tournaments being pretty abbreviated compared to last year, you’re almost definitely going to hit anything that’s even remotely popular. There’s sort of a line there — what is too obscure to be ignored for a tournament like this and what isn’t? Obviously based on the previous tournament, Smeargle is not so obscure that it can be ignored, and I think if you want a bunch of points for the next one, you’ve gotta have a good plan for it. I think it’s still important to figure out what you can ignore though, and what you can afford to take weaker matchups against. Perish Trap? Harder Trick Room? Rain? Sun? Smeargle? You know you’re going to face an abundance of goodstuffs offensive crap around Kangaskhan/Talonflame and apparently Smeargle, you know you’re gonna see a lot of suddenly threatening badstuffs offense with Greninja/Talonflame/(mega), what aren’t you going to see much of? How many times are you willing to play a bad matchup to sacrifice it to improve your matchups against more common archetypes?
     
    I think this issue is more important this year than any other year.  In 2012 and 2013 we had these defensive behemoths (Cresselia, Landorus-T) and such good speed control (Cresselia, Thundurus) that usually made your bad matchups less outrageous.  This year most of the Pokemon are flimsier, so it can be harder to just bulk out a bad matchup and rely on your playing to make up the distance.  Therefore, I think the players that survive Swiss this year are going to be even more dependent on their ability to survey the metagame than in other years.
     
    As far as the frequency line for bad matchups, I’m gonna go out on a limb and say you shouldn’t have rough matchups against any combination that appears more than 8 percent of the time.  If we have a 12 round Swiss, then realistically you can go 9-3 and have a shot, but let’s say you don’t want to leave things to chance, so 10-2.  This means you can drop a game due to a crappy matchup and then have another game to get outplayed or have some bad luck.  I think right now it’s hard to gauge combinations that appear more than 8 percent of the time based on the stats we have, but I guess we’ll have to make do with our season 3 Battle Spot stats in addition to whatever we get from Showdown.

    Responses in bold since quoting multiple times doesn’t go with lazy bearsfan.
     

    Some good thoughts here, and some pretty decent advice on not having hard autolosses just as we are about to get the June IC underway.
     
    Regarding matchmaking… from what I’ve heard, it tries to pair you up with someone relatively close to your ranking first, but if nothing bites the fishing line after a while, it will broaden the matchmaking range (and will cancel out if it still can’t find anything). Because of this, it can be a bad thing to try to get too far ahead of the pack too quickly, as you may end up constantly being unable to find people with ratings closer to where you are. While things will always depend on who’s online and where they are at, it seems like the tourney either encourages people to pace themselve and spread out their battless a bit, or to go on a 60 battle marathon on the last couple of days if they want a better shot at a good rating.
     
    I also kind of doubt a 90% Win Rate is all that necessary for T4. I think I was 4th last time by a few points and my W/L was closer to 80% than 90%.

    I guess what I’m advocating for is a smaller tolerance before it cancels out altogether.  No comment on the pacing of battles since I haven’t developed any particular thoughts or feelings on the subject yet.
     
    Also the 90% win rate was just to make a hypothetical case.  Turns out it was significantly lower, but I made the point clear enough.  Sometimes I get wrapped up in particulars.
     

    “I’m going to arbitrarily and hypothetically say that a 90 percent win-loss ratio is necessary to reach into the top 4.”

    I collected a bit of data regarding this after the tournament.

    4th – R Inanimate – 1874 (41-12) [77.4%]
    9th – joej m – 1865 (46-14) [76.7%]
    20th – Birch – 1832 (35-10) [77.8%]
    26th – Primitive – 1827 (48-12) [80.0%]
    31st – bearsfan – 1821 (35-9) [79.6%]
    50th – LPFan – 1809 (36-10) [78.3%]
    65th – Mrbopper – 1804 (40-16) [71.4%]
    69th – kingofmars – 1802 (43-16-1) [72.9% excluding the tie]

    A interesting point regarding the Autolosses section of your article is you actually ended up with a very high win percentage. Although at the same time it can be seen that your score seemed to suffer from what I would assume are big hits in your losses.

    One topic that wasn’t address in the article that I haven’t heard much people talk about is the CP payout. Right now Wi-fi is worth less than a Premier Challenge. The reasons for the low payout was obvious last year due to legitimacy issue. However, with some of the larger options fixed, should they now be worth more and how much?

    Yeah I got burned big on my losses, which is what caused me to write about the rating disparity thing.  One of my losses was to a guy about 100 points lower than me.  If s/he was in my rating neighborhood, I would just be like “eh, it’s Pokemon.”  Considering s/he was so much lower, I was like “eh, it’s Poke- wait crap my rating!”, and I think it hurt my overall ranking.
     
    I definitely think that the CP payout should be adjusted to be more than a Premier Challenge.  We have a much wider pool of players vying for a smaller pool of CP.  To me, it takes way more skill to do well in an International Challenge than win a Premiere Challenge (exceptions to hotspot areas like NY, California, etc.).  The reward should reflect that.

  8. Yveltal says:

    I think the match limit is a very positive thing because it prevents giving an unfair advantage to those who have more time, or whose teams play faster. Think about this for children in particular, or people with lives during the tournament: you shouldn’t be handicapped in the sense of being unable to overcome a bad loss, just because you can’t play 30, 40, 50 matches in a day. Having no limit is just unhealthy, and of course some people can and will go crazy, but many more will not — can not — and will rightfully feel underprivileged. Others will feel obligated to play as much as they can, which just destroys the fun of it because it is a stressful chore to play that many matches in a day trying to overcome one loss, which may turn into two, or three. It’s not healthy, it’s not fun, and it’s not fair to those who can’t stretch their waking lives to out-play their more gungho counterparts.

  9. cobalte says:

    Seconding for more ICs spread throughout the year (hopefully not just placed in a bundle around Nats time)
     
    I’m not sure how to prove it but I also had some problems with having players disconnect on me and then when coming back on to play sometime later the disconnect win was no longer there and my rating was lower. Which was pretty frustrating.

  10. Tief says:

    The problem you state with the ELO like rating system is a problem inherent with any game that has chance.

    ELO was designed for a pure skill game (chess). As such anyone that is 100 higher in rating is pretty much not going to lose unless the other player is underrated for some reason (perhaps just starting out).

    I think the only way to keep this from happening would be if the Wifi tournaments actually kept track of a swiss record. This doesn’t fix everything as a lucky streak can catapult a lesser player up quite high.

    Both systems will have problems. If every match was a BO3 then it would further add a skill level factor, but would increase the problem of disconnects and the time limit problems.

    The fact is that there will never be a situation where a highly skilled player is guaranteed a win over a lesser player because there is that luck factor built into the game.

  11. Tief says:

    The problem you state with the ELO like rating system is a problem inherent with any game that has chance.
    ELO was designed for a pure skill game (chess). As such anyone that is 100 higher in rating is pretty much not going to lose unless the other player is underrated for some reason (perhaps just starting out).
     
    I think the only way to keep this from happening would be if the Wifi tournaments actually kept track of a swiss record. This doesn’t fix everything as a lucky streak can catapult a lesser player up quite high.
     
    Both systems will have problems. If every match was a BO3 then it would further add a skill level factor, but would increase the problem of disconnects and the time limit problems.
     
    The fact is that there will never be a situation where a highly skilled player is guaranteed a win over a lesser player because there is that luck factor built into the game.

  12. The problem you state with the ELO like rating system is a problem inherent with any game that has chance.
    ELO was designed for a pure skill game (chess). As such anyone that is 100 higher in rating is pretty much not going to lose unless the other player is underrated for some reason (perhaps just starting out).

    Maybe you should learn what ELO is before posting. If one player had a true rating of 100 elo higher than the other it would be equivalent to a 64% chance of winning. “Pretty much not going to lose” is not the case at all.

    Also on ELO theory, not finishing your 60 games hurts your ELO. Unless you think you don’t deserve your rating, not finishing your games is the equivalent of intentionally drawing your remaining games. If birch played out his games with the same skill previously used he would have finished at 1942 and he wouldn’t have let America down by letting the Japanese win.

    If your average opponent at the time of playing them was 112 points lower than you and you managed to win 90% of your games your rating would be around 1968. I think in previous tournaments win percentages were higher because there wasn’t a rating window filtering many matches against sub 1500 players. (For reference Randy’s average opponent rating was about 35 less than his for each game, and he had a 77% win percentage)

  13. Tief says:

    Oh no…my random numbers that I typed in my 5 minute post doesn’t quite fit. How about a 400 rating difference which means the higher player has ~91% chance to win?
    The point is still valid. ELO is based around a pure skill game and there is more randomness introduced due to the mechanics of the game (I don’t want to open up that can of worms, I’m fine with the chance factors inherent to Pokemon). This can and does change that 91% chance to win.
     
    Using an ELO like system with any game that has a chance factor isn’t perfect because a string of bad luck takes the skill factor out of the game imbalancing the ratings leading to “…players sitting around 1750 flipping tables because they had a streak of unlucky rolls against a 1600 player…” This was my point 
     
     
    I don’t mind looking like an idiot. Hey, I made the post quick and sloppy. My question is, were you actually trying to be helpful or merely trying to “put this noob in his place”?

  14. Also on ELO theory, not finishing your 60 games hurts your ELO. Unless you think you don’t deserve your rating, not finishing your games is the equivalent of intentionally drawing your remaining games. If birch played out his games with the same skill previously used he would have finished at 1942 and he wouldn’t have let America down by letting the Japanese win.

     

    I wont let you guys down again. :(

Leave a Reply

Back to Top ↑