Why is online in this game so bad?

Status
Not open for further replies.

Xquirtle

Smash Journeyman
Joined
Jan 30, 2015
Messages
232
Location
Columbus, Ohio
I suspect the player pool for competitive free for alls is substantially lower than that for 1v1 ranked matchmaking queuing.




What positive experiences did you have with Smash 4's For Glory Mode in regards to matchmaking?

How were they different than the current matchmaking system?

I'm not talking about being paired against players with differing rulesets. I think we can all agree that's annoying, and that most players would prefer to have a longer queue time or be paired with less ideal opponents than they would to be paired against someone with a ruleset they deem unacceptable.

I am discussing strictly match-making. I would suspect the caveats of the system to be similar, though I never played Smash 4 online.
the matchmaking was much worse in smash 4, and you had very little information on how well you are doing other than your win rate. Like you could see total games played / won for your tag, but very little to know where you stand overall. There was no displayed ranking or rating etc and you could just create a new tag to wipe all of the stats out. Ultimate is noticeably better at actual matchmaking, although, i would not call it good. Its just obvious that it is at least trying to match based on some sort of rating that is not specifically GSP.

The only thing that smash 4 has on ultimate is that it had the decency to separate out the joke rule sets into a different game mode, and provided us with For Glory, that was purely 2 stock 1v1 FD no items. Everything in ultimate should also be this way. 1v1 and 2v2 serious modes, plus an open rule set mode where you can do whatever you want. Now, i think that they combined them so that there will be enough players to quickly fill the joke rule set games. If it were separate, nobody would ever queue those, because everybody hates them. Or... at least, the wait times would be a good bit longer.
 
Joined
Apr 28, 2008
Messages
5,921
Location
California
You're playing at the very top of the bracket. The game does not have an infinite amount of people queuing. It has to match you against players that are both close to you in rating and close to you geographically. The opponent it finds is a compromise of the two.

Given you're at the point where you're gaining few points for wins and taking big hits for losses, it means you have a high win / loss ratio, which means the game is incapable of finding you an opponent it deems both reasonably close geographically and close to you in skill level.

Elite Smash, by definition, is an Elo barrier defined as the upper 3.5% of players. That alone, despite what many say otherwise on these forums, is a very exclusive pool of individuals.

Here's a graphic describing this, below.

You are not just in Elite Smash. I just started hitting ES, and I get about 30-50k per win. You're getting a few k per win or less. You are probably deep in Elite Smash, and would probably destroy me online in a set. I'd likely lose to you 3 out of every 4 matches or worse, so you're absolutely at least 1, probably 2 or more standard deviations into Elite Smash.

Elo ratings, by definition, mean that one standard deviation of Elo represents 200 rating. Players that differ by 200 Elo rating, by definition, will have a 3:1 win loss rate against one another.

If the game is incapable of finding players around your Elo rating due to just how high it is, it will start pairing you against players lower than you. If you're fighting an opponent who is one standard deviation lower than you, you will beat them 3 times out of every 4, and you will lose 3x the points for each loss against this person. If both of your Elo ratings are a true reflection of your skill, you will, by the very definition of the Elo system, go 3-1 against each other, and net 0 points long-term as you perform an infinite number of matches. If your skill goes up and you can prove it, you'll go 4-1, and, at infinity, you'll gain a few points until the win / loss ratio reflects your proper Elo ratings, and you'll end up with a higher Elo score (GSP is a proxy for this), and you'll start losing 4x the points you gain against this person as your new Elo rating stabilizes.

This is going to be an issue with any matchmaking system. Players at the very top of the pool just are very few and far between. Your rating, by definition and by how Elo systems work, is high enough to where you don't get many more points pummeling the same players that are one or more standard deviations below you in Elo. When you occasionally do lose to them, you take a huge hit. If you didn't, your rating would inflate to infinity and the entire system would break down. No matter how big a player pool is, there will be people at the upper tip of the distribution who cannot often be paired against other players of similar Elo ratings.
View attachment 235378
This is pretty interesting. My GSP is currently 5.364 million with Dark Samus, but my Samus is higher than that. I don't remember the exact numbers. I've lost GSP so I'm working my way back up with Dark Samus so she's right there with Samus. I don't remember the last three digits for my GSP, but the Elite GSP site says my GSP is higher than 98.74% of online players. No one is at 5.4 million GSP, right?
 
Last edited:

Predatoria

Smash Journeyman
Joined
Jan 30, 2019
Messages
325
Location
Raleigh, North Carolina
Switch FC
SW-5219-6817-7975
This is pretty interesting. My GSP is currently 5.364 million with Dark Samus, but my Samus is higher than that. I don't remember the exact numbers. I've lost GSP so I'm working my way back up with Dark Samus so she's right there with Samus. I don't remember the last three digits for my GSP, but the Elite GSP site says my GSP is higher than 98.74% of online players. No one is at 5.4 million GSP, right?

I .... had no idea you could find that information. This is a wealth of a finding for what I want to do here.

I can do so much with this! Here's a bit of mathematical poking around on my end demystifying what GSP actually translates to:

Here's a bit of MATLAB code to take GSP and translate it into effective Elo ratings and, thus, win rates against other players:

%These two data values are from elitegsp.com
MaxGSP = 5425647 / 1000000;

mu = 1500;
sigma = 285.71;
ELOs = 1500:sigma/2:1500+5*sigma;
ELOs = [1500-5*sigma:sigma/2:1500-1*sigma/2 ELOs];

CDF = 0.5 * (1 + erf((ELOs - mu)/sqrt(2)/sigma));

GSPs = MaxGSP * CDF;



figure(1);
scatter(GSPs, ELOs);
text(GSPs+0.1, ELOs, strcat(string(round(GSPs, 3)), 'M GSP,', string(round(ELOs, 0)), ' Elo' ));
text(GSPs-0.4, ELOs, strcat(string(round((ELOs - mu)/sigma, 1)), '\sigma'), 'fontsize', 20);
hold all;
plot(GSPs, ELOs);
line([0 6], [2000 2000], 'linewidth', 3, 'color', [1 0 0]);
text(2, 2100, strcat('ELITE SMASH = 1.8\sigma'), 'fontsize', 20, 'color', [1 0 0]);
xlabel('GSP (Millions)');
grid on;
ylabel('ELO Rating');
title('ELO Rating from Smash Ultimate GSP', 'fontsize', 25);

figure(2);
sigmas = -5:0.5:5;
winrate = 50 * (1 + erf(sigmas/sqrt(2)));
plot(sigmas, winrate);
hold all;
scatter(sigmas, winrate);
grid on;
title('Win Rate as a Function of Standard Deviations', 'fontsize', 25);
xlabel('Standard Deviations (\sigma) of Difference between Players');
ylabel('Win Rate (Percent)');
The above, simple MATLAB code generates these two plots, which I think people here will find tremendously informative:
Plot1.jpg



Plot2.jpg




Here's an example of the usage of these figures. My GSP is 5.316 Million currently (on Ridley). This means my Elo rating with Ridley is somewhere around 2150, or around 2.25 sigma. You said your GSP on Dark Samus is 5.364 Million, which puts you at about 2200, just by eyeballing the plot, or about 2.40 sigma. I may generate math functions that find these exact values later, as right now I'm really just eyeballing the plots to get these numbers.

We differ by 0.15 sigma. This means, according to the game, if my Ridley fought your Dark Samus, it would expect me to beat you about 45% of the time. If we were to match up 20 times, and I won 9, that is the expected win rate for my Elo rating versus yours. We would also gain and lose GSP values that would put us right back where we started after these 20 matches, if I were to win 9 of them.

If, however, I won 10, my GSP would go up slightly and yours would fall slightly. This is because I did better than my Elo rating predicted I would against you, and I have defied the Elo rating's expectations of my performance.

I may also come back to this later and generate some functions that calculate how much GSP you'll win or lose, based on your and your opponents GSPs. Now that I can extract elo ratings and sigma differences for arbitrary GSP values, doing so won't be too difficult.

Also, notice how tightly the GSP points get at both ends of the plot. That's because what you're actually doing behind the scenes is gaining or losing Elo rating, and the GSP is just mapped to the cumulative distribution function that is your GSP divided by the maximum GSP (max GSP is just how many participants there are).

Another interesting observation is just how many sigmas you can go through at the upper end of the bracket. Players such as Zero or Tweek would probably be way up there at 4 sigma, 5 sigma, or beyond. A regular Elite Smash player may be around 2 sigma or so, which is about where I am right now. That means if I were to fight Tweek, there would potentially be a 2-3 sigma difference in our skill level. With three sigmas of difference (look at the second plot), my chances of winning against him are less than 1 percent. I'd get wreckt, but he'd gain virtually 0 elo rating for beating me.
 
Last edited:

DeDeDIke

Smash Apprentice
Joined
Jul 4, 2019
Messages
87
Location
Vancouver
Pretty interesting graph you got there. Probably many players here on the board would find this helpful. Judging from your graph, I guess I'm at about 1.5 sigma and ~1960 ELO rating with my highest GSP main. I don't see many ppl above 2.5 sigma (or ELO rating over 2200) online, but can somewhat put up a good fight against players in the 2 sigma range. There are some ppl I've seen at 5.3 mil GSP who are still not in Elite Smash (which is weird to me, since that high of a GSP SHOULD mean you're in ES), and I fight against some of them pretty often. I think almost everyone here (correct me if I'm wrong) would get bodied by someone who is at 2.5 sigma (or close to it). Having a GSP calculator would be nice. Also, I'm assuming these plots are strictly for 1 v 1s only; how would GSP gain/loss be calculated in 3 player/4 player FFA or team battles?
 

Sucumbio

Smash Powerslave
Moderator
Joined
Oct 7, 2008
Messages
5,244
For Glory had default 2 min no items. Probably 90 percent of the 74k or so matches I logged were against people who were former brawl players who adapted to 4s mechanics at about the same pace or were new but quick enough to keep up with the rest of us. The other 10 percent were spammers, teabaggers or taunt enthusiasts. Because the glory mode was SO specific it attracted only specific opponents after the game was out for a year. And right up until Ultimate came out it was fairly dependable for a good time.

The lag issues were much better than Brawl as was the player quality since for glory was basically the quickest way to play competitively and so attracted those types of players.

Ultimate has undone this by shifting the importance of playing competitively to maintaining a certain GSP. Since GSP is calculated without certain issues in mind you're likely to languish in lower threshold ranges unless you prefer 1v1 which is ES's predominant ruleset.

These issues include lag matches which become unwinnable, sore losers dropping out mid match (no GSP gained even if you win), being matched to players way above or below your skill, items randomly available, bad stages randomly available etc. Without the glory ruleset being uniform there's no consistency in match type or opponent. Pretty much everything goes just like in Brawl days.

Since battle arena is unranked it means that aspect of the game is moot. But some players look for being able to say they're not only better than others but that they have a rank to prove it. It's fair to want that but quickplay makes it randomly impossible without playing only 1v1. So my biggest gripe is the loss of team battle glory mode.

Tournament mode is great this time around. By organizing the pools for us we aren't stuck waiting forever to get people to join our custom tourneys (4s allowed you to make your rules so it was the true predecessor to battle arenas). My only wish is that they make a no items 2 stock 3 min 1v1 or teams all the time.

Or make a mode like quickplay that has that.

Basically bring back for glory lol.
 

Predatoria

Smash Journeyman
Joined
Jan 30, 2019
Messages
325
Location
Raleigh, North Carolina
Switch FC
SW-5219-6817-7975
Pretty interesting graph you got there. Probably many players here on the board would find this helpful. Judging from your graph, I guess I'm at about 1.5 sigma and ~1960 ELO rating with my highest GSP main. I don't see many ppl above 2.5 sigma (or ELO rating over 2200) online, but can somewhat put up a good fight against players in the 2 sigma range. There are some ppl I've seen at 5.3 mil GSP who are still not in Elite Smash (which is weird to me, since that high of a GSP SHOULD mean you're in ES), and I fight against some of them pretty often. I think almost everyone here (correct me if I'm wrong) would get bodied by someone who is at 2.5 sigma (or close to it). Having a GSP calculator would be nice. Also, I'm assuming these plots are strictly for 1 v 1s only; how would GSP gain/loss be calculated in 3 player/4 player FFA or team battles?

I have heard the game does, on occasion, pair people up across the Elite Smash border. If you're queuing FFA's, it likely has a smaller player pool and is bringing ES members into your matches more frequently out of necessity. I don't think I've seen that happen to me yet, but I always queue 1v1 'Ranked Matchmaking.' I believe the cutoff for ES is somewhere just after 5.2 Million right now, as I remember being at exactly 5.200M and not having it, and then winning a game and getting it with 5.211M just last Wednesday.

You'd stand a reasonable chance at winning against somebody who was 200 or so Elo higher (2.5 sigma, compared to your 1.5), as that would suggest you'd win about one match out of every four. I've fought people at weeklies who beat me most matches, and usually I can at least get them down to their last stock and put up a good fight. I'll snag a win occasionally. They're probably about a sigma of deviation more skilled than I am. If I, however, get paired against a PR player in bracket, I usually get, as I like to call it, 'promptly 3-stocked.' They're probably multiple sigmas more skilled, and my chances of winning are very small.

I may create a true GSP to Elo calculator, and a calculator where you can put in your GSP, their GSP, and how much you'd gain or lose for a win or loss against that opponent. It'd certainly help demystify what's going on behind the scenes here with Nintendo's Smash GSP code. It really is just an Elo system hiding behind a proxy rating, 'GSP.' What I don't like about GSP is that it distorts heavily at the tail ends, compressing many standard deviations of players into one very similar sounding number.

I'm not too sure how it treats free for alls. I've never done them with GSP involved, and typically only do FFAs with buddies in real life. I think it does some odd thing where first place is a full win, second place is half a win, third place is half a loss, and fourth place is a full loss, but I don't know what Elo it's comparing yours against to compute the gains or losses. It may just be treating your 1st place win as 1/3 a win against each player and taking their relative Elo ratings into account like that. That would ensure the net gains / losses for everyone's rating sum to zero, which is critically important to avoid 'rating leaks.'

As for team battles, I'm not sure. I queue ranked matchmaking 2v2s sometimes with friends, and I have no idea what it's doing. My GSP never changes. It seems like the opponents I'm paired against are truly random, and I don't think matchmaking seems to be implemented here, but I could be wrong. I pay less attention to it than I do the 1v1s.
 
Last edited:

Sucumbio

Smash Powerslave
Moderator
Joined
Oct 7, 2008
Messages
5,244
Afaik or at least have observed in a 4 player match ffa winner gets largest gsp gain 2nd place small gain until you're closer to the threshold then neither gains or loses 3rd loses some and 4th the most. In teams your partner used to be static but since 3.0 it's randomized. Winning team gets gsp gain losers lose.
 
Joined
Apr 28, 2008
Messages
5,921
Location
California
I .... had no idea you could find that information. This is a wealth of a finding for what I want to do here.

I can do so much with this! Here's a bit of mathematical poking around on my end demystifying what GSP actually translates to:

Here's a bit of MATLAB code to take GSP and translate it into effective Elo ratings and, thus, win rates against other players:



The above, simple MATLAB code generates these two plots, which I think people here will find tremendously informative:
View attachment 235486


View attachment 235487



Here's an example of the usage of these figures. My GSP is 5.316 Million currently (on Ridley). This means my Elo rating with Ridley is somewhere around 2150, or around 2.25 sigma. You said your GSP on Dark Samus is 5.364 Million, which puts you at about 2200, just by eyeballing the plot, or about 2.40 sigma. I may generate math functions that find these exact values later, as right now I'm really just eyeballing the plots to get these numbers.

We differ by 0.15 sigma. This means, according to the game, if my Ridley fought your Dark Samus, it would expect me to beat you about 45% of the time. If we were to match up 20 times, and I won 9, that is the expected win rate for my Elo rating versus yours. We would also gain and lose GSP values that would put us right back where we started after these 20 matches, if I were to win 9 of them.

If, however, I won 10, my GSP would go up slightly and yours would fall slightly. This is because I did better than my Elo rating predicted I would against you, and I have defied the Elo rating's expectations of my performance.

I may also come back to this later and generate some functions that calculate how much GSP you'll win or lose, based on your and your opponents GSPs. Now that I can extract elo ratings and sigma differences for arbitrary GSP values, doing so won't be too difficult.

Also, notice how tightly the GSP points get at both ends of the plot. That's because what you're actually doing behind the scenes is gaining or losing Elo rating, and the GSP is just mapped to the cumulative distribution function that is your GSP divided by the maximum GSP (max GSP is just how many participants there are).

Another interesting observation is just how many sigmas you can go through at the upper end of the bracket. Players such as Zero or Tweek would probably be way up there at 4 sigma, 5 sigma, or beyond. A regular Elite Smash player may be around 2 sigma or so, which is about where I am right now. That means if I were to fight Tweek, there would potentially be a 2-3 sigma difference in our skill level. With three sigmas of difference (look at the second plot), my chances of winning against him are less than 1 percent. I'd get wreckt, but he'd gain virtually 0 elo rating for beating me.
I wonder if that's the reason why the Samus/Dark Samus main, YB, doesn't gain any GSP. He currently has the highest GSP for both Samus and Dark Samus, and when I've watched him play on live stream, he wouldn't gain any points. I figured he reached the limit, since GSP has a cap and won't go any higher until at a later time. It also seems that some characters can't reach higher than others in terms of GSP. YB's Dark Samus never passes Samus' GSP. I haven't seen him around lately and the last thing I saw him doing was playing in the online tournament that Nintendo recently added to SSBU. I've tried adding YB's GSP to the Elite GSP site, but it won't let me for some reason, even though I see lower values only being posted up.

By the way, is the Elo ranking system a good ranking system, or are there ranking systems that are better? If the Elo ranking system is a good indicator of skill, I may just have to put up with it in spite of my disdain for it. I really appreciate all this information you've gathered. I don't know how you've done it, but it's really neat. Thank you.
 
Last edited:

Xquirtle

Smash Journeyman
Joined
Jan 30, 2015
Messages
232
Location
Columbus, Ohio
I .... had no idea you could find that information. This is a wealth of a finding for what I want to do here.

I can do so much with this! Here's a bit of mathematical poking around on my end demystifying what GSP actually translates to:

Here's a bit of MATLAB code to take GSP and translate it into effective Elo ratings and, thus, win rates against other players:



The above, simple MATLAB code generates these two plots, which I think people here will find tremendously informative:
View attachment 235486


View attachment 235487



Here's an example of the usage of these figures. My GSP is 5.316 Million currently (on Ridley). This means my Elo rating with Ridley is somewhere around 2150, or around 2.25 sigma. You said your GSP on Dark Samus is 5.364 Million, which puts you at about 2200, just by eyeballing the plot, or about 2.40 sigma. I may generate math functions that find these exact values later, as right now I'm really just eyeballing the plots to get these numbers.

We differ by 0.15 sigma. This means, according to the game, if my Ridley fought your Dark Samus, it would expect me to beat you about 45% of the time. If we were to match up 20 times, and I won 9, that is the expected win rate for my Elo rating versus yours. We would also gain and lose GSP values that would put us right back where we started after these 20 matches, if I were to win 9 of them.

If, however, I won 10, my GSP would go up slightly and yours would fall slightly. This is because I did better than my Elo rating predicted I would against you, and I have defied the Elo rating's expectations of my performance.

I may also come back to this later and generate some functions that calculate how much GSP you'll win or lose, based on your and your opponents GSPs. Now that I can extract elo ratings and sigma differences for arbitrary GSP values, doing so won't be too difficult.

Also, notice how tightly the GSP points get at both ends of the plot. That's because what you're actually doing behind the scenes is gaining or losing Elo rating, and the GSP is just mapped to the cumulative distribution function that is your GSP divided by the maximum GSP (max GSP is just how many participants there are).

Another interesting observation is just how many sigmas you can go through at the upper end of the bracket. Players such as Zero or Tweek would probably be way up there at 4 sigma, 5 sigma, or beyond. A regular Elite Smash player may be around 2 sigma or so, which is about where I am right now. That means if I were to fight Tweek, there would potentially be a 2-3 sigma difference in our skill level. With three sigmas of difference (look at the second plot), my chances of winning against him are less than 1 percent. I'd get wreckt, but he'd gain virtually 0 elo rating for beating me.
I think that this analysis holds true on either tail end of the GSP spectrum, where all of the players have a large number of games played, but not so well in the center. We can assume that some large percentage, probably 30-50%, of character specific GSPs are just fake rankings on characters that players have never even played a single game on. This stacks a bunch of trash data all around 1500 ELO, + or - some slight adjustment for some system made up by ninetendo. These players aren't participating in the exchange of ELO, but they are included in the straight ranking of players that is called GSP.

An extreme example of this would be if GSP included every single person on earth. You would have billions of people at ~1500 ELO that don't even play the game. They never win or lose a game, BUT if you lose your very first game, so 0-1 on a character, you will drop below them on the gsp straight ranking, even though you are a better player since you actually play the game. You would suddenly be at sub 500 ELO as projected by the above chart, however, you've only lost one game, so you should really be around 1486 ELO. Regardless, we have a less extreme version of this, but its still a huge portion of the GSP ladder.

I think GSP is just such a trash system that, unfortunately, you cannot apply analyses depending on normality very effectively. It doesn't provide strong correlations to, well, anything, due to the huge injection of 'fake' data. That being said, it probably fits the curve perfectly once you break out of the center that is ass jacked up with fake data.
 

DeDeDIke

Smash Apprentice
Joined
Jul 4, 2019
Messages
87
Location
Vancouver
I may create a true GSP to Elo calculator, and a calculator where you can put in your GSP, their GSP, and how much you'd gain or lose for a win or loss against that opponent. It'd certainly help demystify what's going on behind the scenes here with Nintendo's Smash GSP code. It really is just an Elo system hiding behind a proxy rating, 'GSP.' What I don't like about GSP is that it distorts heavily at the tail ends, compressing many standard deviations of players into one very similar sounding number.
Are there any factors that might affect the amount you gain/lose? For example, your win rate in the last 10 games (or last 50 games) or the total number of games you've played on a certain character. I've noticed that when I go on a losing streak (sometimes I lose ~150k GSP before managing to go back up, since I only earn 7k to 9k per win on a daily basis), and then I win a match, it gives me more GSP than I regularly would (~17k as compared to the normal 8k).
 

TheDuke54

Smash Journeyman
Joined
Apr 19, 2016
Messages
382
That isn't the issue though. We are all glad that Smash is online and we can play with randoms and friends. The problem is we are paying a subscription fee for features that were once far better and for free. And Nintendo has been delivering substandard services for online thus far and ignoring most complaints. At least as far as online random play is involved.
 

Sucumbio

Smash Powerslave
Moderator
Joined
Oct 7, 2008
Messages
5,244
The more matches you play with a given character the less your GSP swings with a loss or win especially if you are towards the elite threshold.

Heh yeah this go around was definitely one step forward two steps back. So long as they keep the no items tournaments going tho I'm satisfied.

And I gotta have dlc cause fe is out now and I plan on getting the dlc pack.
 

Xquirtle

Smash Journeyman
Joined
Jan 30, 2015
Messages
232
Location
Columbus, Ohio
The more matches you play with a given character the less your GSP swings with a loss or win especially if you are towards the elite threshold.

Heh yeah this go around was definitely one step forward two steps back. So long as they keep the no items tournaments going tho I'm satisfied.

And I gotta have dlc cause fe is out now and I plan on getting the dlc pack.
Seems to be two separate things from my experience. you have the massive GSP jumps and losses right when you start a fresh character, and then once that stabilizes, it also changes at different rates depending on where you're at on the ladder. much lower gains and losses if youre at the very bottom or top.

super jank is one way to put it lol.
 

Predatoria

Smash Journeyman
Joined
Jan 30, 2019
Messages
325
Location
Raleigh, North Carolina
Switch FC
SW-5219-6817-7975
I wonder if that's the reason why the Samus/Dark Samus main, YB, doesn't gain any GSP. He currently has the highest GSP for both Samus and Dark Samus, and when I've watched him play on live stream, he wouldn't gain any points. I figured he reached the limit, since GSP has a cap and won't go any higher until at a later time. It also seems that some characters can't reach higher than others in terms of GSP. YB's Dark Samus never passes Samus' GSP. I haven't seen him around lately and the last thing I saw him doing was playing in the online tournament that Nintendo recently added to SSBU. I've tried adding YB's GSP to the Elite GSP site, but it won't let me for some reason, even though I see lower values only being posted up.

By the way, is the Elo ranking system a good ranking system, or are there ranking systems that are better? If the Elo ranking system is a good indicator of skill, I may just have to put up with it in spite of my disdain for it. I really appreciate all this information you've gathered. I don't know how you've done it, but it's really neat. Thank you.
I'm glad you enjoyed reading all my math attacks! Most people back away slowly when I start talking about game mechanics, code, math, or anything like this so I'm thrilled you found it to be informative and interesting.

I got a bit weary typing up these long posts over the weekend, and my MATLAB script was actually at work.

Elo is by far the most commonly used ranking system for video games, though many games alter it in some way, disguise it, or make their own modifications to the system. Smash, for example, has the arbitrarily defined 'Elite Smash' threshold, does some funky stuff for determining your Elo on characters you've never even played before, and has its own system in place to handle Elo exchanges in FFAs, all of which go beyond the scope of the original Elo ranking system, which was used for Chess matches.

On the topic of the actual post though, I've been finding Ranked Matchmaking 1v1s to be a great way to practice lately when I can't find rl friends to play with. I mostly just go in there, ignore gsp, ignore winning or losing, and try to sharpen up my game. Sometimes I'll even SD out of a match if it's really laggy or has a ruleset I find intolerable, just because I don't want to waste valuable practice time dealing with it, and don't care about taking the GSP hit.
 
Last edited:

TheDuke54

Smash Journeyman
Joined
Apr 19, 2016
Messages
382
Watch out doing that too much though. You'll get banned. If the lag is horrible or the rules just outright suck and I don't want to deal with the match I will also occasionally SD or let the opponent frag me once so it's not considered a SD when I jump off the ledge.

That's another thing that blows. You can't leave the lobby. Even if the QP game takes a minute or so to commence for whatever reason. Why they thought this was a good idea when it wasn't implemented in 4 is beyond me.
 

J.I.L

Banned via Administration
Joined
Apr 30, 2019
Messages
327
I am starting to despise battle arena players. I’m either getting kicked after matches or barred from enterting open arenas (which means I’m getting blocked). My goodness, are people who play smash insecure poorly trained divas? Socially inept freaks who cant take playstyles in a game? Are people just children. Like it’s becoming ridiculous. I don’t have a voice chat so I people dont hear me... I’m just getting kicked left and right. People on battle arenas (I think they represent online smash players in general) are...lol. Talk about People who cannot survive in the real world. Bunch of weak and pathetic babies who need their safe spaces on arena. Why would Nintendo give children the ability to kick people? Idiots. This “kick option” is a joke that should be disabled. To many low lives abuse it.

Edit: I just got kicked again. It’s like 70% of the time I enter a room, I’m kicked after lie one game. i don’t even feel that bad, it’s more comical how degenerativly tribal young people are. Stone Age tribal. Folk. And I’m young myself. And it’s not bad WiFi, the game runs smoothly. It’s just people can’t handle certain playstyles. It’s so shameful. And I just look down on it, honestly. It’s comically pathetic, but depressing how people act like this. Believing in Liberty is not inate with many people. People like dictators or oppressive mob thought. That’s the only thing I can conclude.

Honestly, what makes it not so depressing is that I know I’m a superior player to all of them. It’s the just the facts. I’m more cabable then them. I could crush most of them 1v1. So i know they are blocking me because they are insecure losers who are low lives inside and out. The only annoyance is that these insecure losers have avenues to showcase their emotional unstable mediocrity at my expense.

Hmph, useless disgraces like them shouldn’t have power.
 
Last edited:

Predatoria

Smash Journeyman
Joined
Jan 30, 2019
Messages
325
Location
Raleigh, North Carolina
Switch FC
SW-5219-6817-7975
Watch out doing that too much though. You'll get banned. If the lag is horrible or the rules just outright suck and I don't want to deal with the match I will also occasionally SD or let the opponent frag me once so it's not considered a SD when I jump off the ledge.

That's another thing that blows. You can't leave the lobby. Even if the QP game takes a minute or so to commence for whatever reason. Why they thought this was a good idea when it wasn't implemented in 4 is beyond me.
Ehhh I’ve been doing this for a while to get out of laggy matches, and only once has a warning shown up about it since I had two super laggy opponents in a row. I’d say I probably only do this 0-1 times per play session.

I am starting to despise battle arena players. I’m either getting kicked after matches or barred from enterting open arenas (which means I’m getting blocked). My goodness, are people who play smash insecure poorly trained divas? Socially inept freaks who cant take playstyles in a game? Are people just children. Like it’s becoming ridiculous. I don’t have a voice chat so I people dont hear me... I’m just getting kicked left and right. People on battle arenas (I think they represent online smash players in general) are...lol. Talk about People who cannot survive in the real world. Bunch of weak and pathetic babies who need their safe spaces on arena. Why would Nintendo give children the ability to kick people? Idiots. This “kick option” is a joke that should be disabled. To many low lives abuse it.

Edit: I just got kicked again. It’s like 70% of the time I enter a room, I’m kicked after lie one game. i don’t even feel that bad, it’s more comical how degenerativly tribal young people are. Stone Age tribal. Folk. And I’m young myself. And it’s not bad WiFi, the game runs smoothly. It’s just people can’t handle certain playstyles. It’s so shameful. And I just look down on it, honestly. It’s comically pathetic, but depressing how people act like this. Believing in Liberty is not inate with many people. People like dictators or oppressive mob thought. That’s the only thing I can conclude.

Honestly, what makes it not so depressing is that I know I’m a superior player to all of them. It’s the just the facts. I’m more cabable then them. I could crush most of them 1v1. So i know they are blocking me because they are insecure losers who are low lives inside and out. The only annoyance is that these insecure losers have avenues to showcase their emotional unstable mediocrity at my expense.

Hmph, useless disgraces like them shouldn’t have power.
No offense dude, but if your in-game behavior is anything like how you behave on these forums, I’m honestly not surprised.
 

J.I.L

Banned via Administration
Joined
Apr 30, 2019
Messages
327
Ehhh I’ve been doing this for a while to get out of laggy matches, and only once has a warning shown up about it since I had two super laggy opponents in a row. I’d say I probably only do this 0-1 times per play session.


No offense dude, but if your in-game behavior is anything like how you behave on these forums, I’m honestly not surprised.
And what’s the problem with my behavior on these forums?
 

Coolboy

Smash Journeyman
Joined
Aug 27, 2018
Messages
326
Location
Netherlands
And what’s the problem with my behavior on these forums?
is that a serious question? i am not sure or you really are this oblivious or you are really just a troll trying to portray the typical annoying smash player,
you overreact on ALOT of things with the hate you have on Sakurai and co, not just that but also talking **** on Smash players you take that to the extreme, and in all honesty..if you get kicked everywhere then i also think it's on you..you clearly admitted multiple times you have a troll playstyle soooooo if you get kicked for it..then either stop trolling oooooor keep trolling with the risk of getting kicked everywhere.
if i trolled everyone in BA i wouldn't find it weird that everyone kicks me out of their arena.
but also you brag alot... ''i am a superior player! i kicked alot of asses today!'' or ''i won alot of online tourneys because i am just that great at the game!''

or your BA talk ''i keep winning with ease though i troll them so hard that they kick me for it but they are the immature ones not me!''
do you know what it is with bragging online? many do it but they sadly fail to proof it with screenshots or anything to proof they are as good as they say, i actually don't believe you are as good as you say since you never seem to actually have screenshots or clips to back it up,
not just you though i don't believe anyone who brags unless they got videos/pics/clips to proof it
 

Sucumbio

Smash Powerslave
Moderator
Joined
Oct 7, 2008
Messages
5,244
Good bump. Thinking this thread has outworn its welcome. We have an online suggestion thread stickied among many other useful discussion. Rant threads which this has been for a while are against tos anyway so... bye.
 
Status
Not open for further replies.
Top