Margin of Victory

[bothanspy 12]

MoV means a player is rewarded based on how he played, not based on the difficulty of his games. I still thinks that a full win  with a 115 MoV against a very good player is more valuable than a full win with a 200 MoV against an average player.

 

The difference between a full win and a modified win already gives a notion of how much the difference in points was. So this aspect is already reflected in the score. MoV only adds a layer in order to get rid of an unpopular SoS IMO. It also dictates players how they should play: more than 12 points difference is not enough anymore. This will affect how people will build and play their squadrons, resulting in less variety IMO.

 

My point is whatever the tiebreaker is, the pairing will always be a factor in the ranking, because that's how the swiss system works. SoS seems unfair to many players because you don't have control on your opponents score, but I think MoV is worse because it doesn't consider who you faced. Winning against a champion - no matter how you play - should be rewarded more than playing "well".

[StevenO 2,996]

 

The 50 free points for a bye is completely arbitrary. I wonder why they chose that value. What's the average per-game MoV fr typical players in the Final Cut, I wonder? 50 points is almost certainly larger than the average, making a first-round bye a distinct statistical advantage. At a minimum it's zero risk.

 

To my knowledge, byes are supposed to be stronger than a regular game win. What would be the point of awarding players byes to higher level events if it automatically put them at a disadvantage in tie-breaking?

 

 

Here I remember hearing may complain that byes are HORRIBLE for your SOS because you're "defeating" someone who can't/hasn't won a single game.  That's pretty devastating.

 

Then you have the "super byes" which treat the bye as the finalist whom you happen to have caused his only loss.  To me that is just as crazy, maybe even more so, than counting the bye as win less.  You go from one extreme to the other.

 

If the MoV method treats you as destroying half your opponent's ships with no losses of your own that sounds like a pretty middle of the road answer.  

[Frydaddy 8]

Holy cow! The amount of misinformation about byes, SOS, and how a swiss tournament is suppose to be ran is mind blowing.

MOV is easier, simpler, understandable, linear, basic, straight forward, elementary, uncomplicated...and that is why people like it.

MOV can be manipulated, rewards smashing newbs, promotes turtling when playing good opponents, punishes you when you play a good game against a good opponent.

I can agree that MOV can work if, all the opponents are the same quality and all opponents are playing to win. But that is what we call a mythical unicorn when it comes to tournament play at the local level.

[Vorpal Sword 14,685]

Holy cow! The amount of misinformation about byes, SOS, and how a swiss tournament is suppose to be ran is mind blowing.

Finally something we can agree on! Your posts do indeed contain a great deal of misinformation about byes, SOS, and how a swiss tournament is suppose to be ran [sic]. Accordingly, I'd suggest that anyone wanting to understand why MOV was introduced and how it works should ignore you entirely.

[Frydaddy 8]

Vorpal:

With MOV, what type of player do you want to get matched up against, to further yourself in the standings. Weak or strong?

With SOS, what type of player do you want to get matched up against, to further yourself in the standings. Weak or strong?

In SOS, When a player drops, do the rounds after they drop count against them? Or does their SOS rating stay the same?

With MOV, are you rewarded for smashing a newb?

With MOV, are you rewarded for turtling against a good player?

With MOV, are you punished for playing a better opponent?

IMHO MOV should be used to calculate SOS. it is a great metric to figure SOS, but a lousy first tie break.

[Vorpal Sword 14,685]

I've answered every single one of these questions already.

With MOV, what type of player do you want to get matched up against, to further yourself in the standings. Weak or strong?

The fact that you're posing this question leads me to believe you don't understand Swiss pairing. Your incentive under MOV is to win your match by the largest margin possible. That means that, among those players in your score group, you would prefer to play the player most likely to perform poorly. But there's an important caveat, which I'll get to in a moment.

With SOS, what type of player do you want to get matched up against, to further yourself in the standings. Weak or strong?

Under SOS, your goal is to win your match by a margin of least 12 points. You want your opponent to go on to win as many games as possible so as to increase your SOS. But the same caveat applies here.

The major flaw in both of these questions is that you're ignoring the role of the pairing method. A Swiss tournament treats your record as a proxy for your skill level, and assigns matches to players who are as close as possible to one another in skill level. So when you talk about being matched against a strong or weak player, what you mean is a player who is strong or weak relative to the other players in your score group, and that places the questions on a very different footing.

Swiss pairing rewards you for playing in excess of the level expected for players in your score group--for being better than it thinks you are. Likewise, it "punishes" you for being worse than the estimate of your skill provided by Swiss pairing. Those aren't moral judgments, of course, but they're important things for the tournament scoring system to "know".

There are two non-exclusive possibilities if you lose a match: you were matched against someone who is stronger than his or her record would indicate, or you were weaker than your record would indicate. The biggest weakness of SOS, in a mathematical sense, is that it provides no direction or information whatsoever about how much stronger or weaker you were; MOV provides that estimate.

What you're perceiving as a flaw is that MOV uses that estimate in both directions simultaneously: in a sense, because it doesn't know whether your opponent was stronger than the typical player in the group or you were weaker, it assumes that both are true. That's unfair because if I win, of course it's because I'm stronger; if I lose, of course it's because my opponent was strong! No one wants the system to assume that his or her opponent was average and he or she just sucked. You think SOS is better because it doesn't provide any estimate at all, just a direction--but that's a flaw in the system, not a strength.

Moreover, here's the last piece of the picture, and it's the dirty little secret of SOS: you don't care how strong your opponent was, as long as he or she is stronger than the players faced by others in your score group. If I'm the only person at 5-0, it doesn't matter if three of those wins were against small children. (That's unlikely, of course, but stick with me for a moment.) If I'm 4-1, all I have to do is hope my opponents had better records on average than the other people who went 4-1. So what I want to play, under SOS, is always the strongest player that (a) I can beat reliably and (b) won't lose more matches than average. You can say that means I want to play strong players under SOS, but that's not true (or at least not complete): it's exactly equivalent to say that I play against a weak player who goes on to be just a bit stronger than each of his or her opponents.

In other words, you (meaning Frydaddy speficially, not a general "you") are making assumptions about "strong" or "weak" that can't be generalized over the whole tournament because they entail assumptions about your opponent as well as all of your opponent's other opponents. What I want isn't to play a strong player, it's for the skill levels and pairings of a lot of other people to align in such a way as to make me look good, and I don't really care how that happens.

MOV, by contrast, does care about your opponent's actual skill level. If you lose by a lot, it says so; if you win by a lot, it says so. In combination with Swiss pairing, it keeps testing the implications of that statement: if you win by a lot, your next match gives you an opportunity to demonstrate whether it was because you were good or because your opponent sucked. If you lose by a lot, your next match gives you a chance to show whether it was because you were terrible or because your opponent was really good. Over the course of an entire tournament, MOV provides more information--far more information, actually, and more accurately on the whole--than SOS does.

In SOS, When a player drops, do the rounds after they drop count against them? Or does their SOS rating stay the same?

In the SOS system outlined by FFG, an opponent who drops from the tournament adds no more points to his or her score. This is precisely equivalent to counting every subsequent match for that player as a loss.

You can substantially improve SOS, at least as regards dropped players, by treating it as a win rate rather than a number of wins, but that's not what FFG told tournament organizers to do.

With MOV, are you rewarded for smashing a newb?

Your question assumes that new players are incompetent and pulls in all the other issues, errors, and bad assumptions I just spent an hour writing out for you. I reject the frame on both counts.

With MOV, are you rewarded for turtling against a good player?

No.

With MOV, are you punished for playing a better opponent?

Unlike SOS, playing a close losing game in MOV is better for you than playing a terrible losing game. That's not a weakness.

[Buhallin 4,520]

I think it's becoming painfully clear that we're not actually discussing the same things here.

 

Most of us are considering the merits of X-wing's SoS-based tournament system and X-wing's MoV-based tournament system.

 

Frydaddy is at least in the ballpark of X-wing's MoV-based tournament system, if an excessively pessimistic view of it...  but I'm not convinced he's ever even read X-wing's SoS-based tournament system, because he has yet to describe anything remotely like it.

 

At six pages in, if someone hasn't bothered to educate themselves to even the most basic familiarity with the topic at hand, they're obviously uninterested in doing so.

[sirhc 82]

 

 

Which is nonsensical, and requires bizarre contrivances like Strength of Schedule & Margin of Victory.

Nobody who loses two Swiss matches in an X-wing tournament is making the final cut in any case, generally speaking.

 

Rhetorical question(s): can you get diced out of chess? In how many other Swiss-style games can you be completely obliterated by bad luck?

 

 

Rhetorical question: How many (X)-2 players made top 8 at a Regional this year? Not counting Australia.

 

Actual Answer:  Kublacon Regional, qualifiers #7 & #8 both had X-2 records, one of which was SprollyG of these boards.

[MajorJuggler 7,752]

One interesting side effect of MoV is that if you do really well early on, it will pair you up against the other players with the same record and a similar MoV. So if you do really well on your first and second game, it is going to become very difficult to keep getting wins, let alone keeping up a high MoV, because you will be playing the best players. In this way MoV does a  better job at quickly "sorting" players than SoS, which is a good thing. But it is brutal if you are the guy at the top.

 

As an example, take the following 2 players. Both are equal skill and both are 2-0. But their MoV is different because the caliber of players that they were randomly matched up with in the first round, and to some extent the 2nd round.

 

Extreme example:

 

Player A is 2-0 and he won both his games with a complete sweep, losing no ships.

Player B is 2-0, and he just barely squeaked by with a win in both rounds, 12 points each time.

 

So:

Player A = 2-0, MoV = 400.

Player B = 2-0, MoV = 224

 

 

Now lets say that there will be 4 more rounds of swiss. After just 2 rounds, who would be projected to make it into the Final Cut? Player A has the better MoV for a tiebreaker, but he will be far more likely to lose a match because he will be playing the #2 player every single game. Player B is going to be theoretically playing weaker opponents, so even though he loses on MoV in a tiebreaker, he is more likely to sweep the rest of his games and go undefeated.

[sirhc 82]

My point is whatever the tiebreaker is, the pairing will always be a factor in the ranking, because that's how the swiss system works. SoS seems unfair to many players because you don't have control on your opponents score, but I think MoV is worse because it doesn't consider who you faced. Winning against a champion - no matter how you play - should be rewarded more than playing "well".

I think most people's challenge with SoS is the impact that other's performance has on how you are ranked at the end of Swiss going into Elimination.  I have seen many a friend that have been negatively impacted by drops and flame-outs.

 

MoV is more fair in that it takes a snap-shot of your performance at a specific point in time which can be used later to rank you after Swiss independent of other player's "snap-shots".  At least I think that's how I want to articulate that... had too much sugar for breakfast. 

Edited July 31, 2014 by sirhc

[VanorDM 11,599]

Good point there MajorJuggler.

The whole point of Swiss and SoS is that you're playing people of similar skill levels. This works great in Chess, because you have a ELO rating and so your skill level is already known. In the first round you'd be paired with people of similar ELO ratings.

In X-Wing there is no ELO rating, so the first round is completely random. You could have someone like Paul or Doug the previous world champs, facing a complete and utter newbie, who picked up a core set, a few other ships and read the rules yesterday... Or you could have two people of similar skill levels face each other.

Because of this, the system they use for Chess doesn't work nearly as well here.

Also as you point out, sweeping the table in the first game, would put you at the top end of the 1-0 group meaning you'll be facing better players in the 2nd and subsequent rounds. That means you're more likely to lose or at least have a much lower MOV for the 2nd round.

[MajorJuggler 7,752]

So, the only question is, who is going to be "that guy" that is the first one to go 6-0 and MoV = 1200? 

 

I will have to start tracking MoV in my stats for kicks.

[Buhallin 4,520]

One interesting side effect of MoV is that if you do really well early on, it will pair you up against the other players with the same record and a similar MoV. So if you do really well on your first and second game, it is going to become very difficult to keep getting wins, let alone keeping up a high MoV, because you will be playing the best players. In this way MoV does a  better job at quickly "sorting" players than SoS, which is a good thing. But it is brutal if you are the guy at the top.

 

This is very true.  One of the negative points against an MoV system is that it can create some perverse incentives...  Back in the SWCCG days, it wasn't uncommon to see someone tank their first game on purpose to try and get easier matchups for the rest of the event.  It was very risky, and I don't think I ever saw it work, but a lot of people don't like that it can work like that.

[VanorDM 11,599]

One of the negative points against an MoV system is that it can create some perverse incentives...

I can see that, you'd want to win but not by too much so that way you're in the bottom half of the group. But then you take a pretty big risk of being knocked out of the final rounds because your MOV is too low.

Seems like the risk is greater then the reward in this case.

[Buhallin 4,520]

 

One of the negative points against an MoV system is that it can create some perverse incentives...

I can see that, you'd want to win but not by too much so that way you're in the bottom half of the group. But then you take a pretty big risk of being knocked out of the final rounds because your MOV is too low.

Seems like the risk is greater then the reward in this case.

 

Nod.  Some players took it as far as actually losing their first game, assuming they could make up the ground.  Obviously if someone went out undefeated they were boned, so it was very risky...  but even the idea that it MIGHT work was unpleasant to a lot of people.

 

Honestly, it's part of why I don't take Frydaddy seriously.  This has always been one of the biggest weaknesses of MoV, in perception if not necessarily reality.  The fact that he's never even brought it up is a good indicator of how little thought he's actually put into this comparison.

Edited July 31, 2014 by Buhallin

[MajorJuggler 7,752]

Ideally, this is what the tie breaking procedure should be. I say ideally because unless you have tournament software this would be difficult to track. The procedures below are the same as used by the NFL when multiple teams are involved, with minor tweaking to make it suitable for X-wing.

http://en.wikipedia.org/wiki/National_Football_League_playoffs

http://www.nfl.com/standings/tiebreakingprocedures

 

After Ranking by Match Points, tie breakers should be:

1. Best Match Points per game in games between the tied players.¹
2. Best Match Points per game in common games between the tied players.²
3. Strength of Victory (Match Points per game of all players that a player has defeated)
4. Strength of Schedule (Match Points per game of all players that a player has played)
5. Best combined ranking of Points Scored and Points Allowed
6. Margin of Victory in common games
7. Margin of Victory in all games
8. Initiative.
9. Coin toss.

 

1 Requires that all players have played at least 1 other tied player within the group at least once

2 Games played against the same opponents

 

• The NFL avoids the issue of "drops" (short seasons) in tiebreakers by using percentages in #1 - #4; translated to X-wing this means using Match Points per Game.
• Strength of Victory is the first tiebreaker that comes up after looking at outright wins by common games. Strength of Schedule then follows.
• Notice how far down Margin of Victory as defined by FFG is on the list. It's number 7 in this system. You'll probably never get that far down. Before you even use MoV you need to first rank all the teams by Points Scored and Points Allowed (Points For and Points Against).

 

That all said, who wants to write a program to use these tiebreakers? 

Edited July 31, 2014 by MajorJuggler

[Vorpal Sword 14,685]

[wall of text]

Now that my blood sugar is back up out of the basement, I'll TL;DR this:

A superficial picture of MOV is that it rewards playing you for playing weaker opponents and penalizes you for playing stronger ones. But

(a) the only thing that makes it different from SOS is that, unlike SOS, MOV provides an estimate of the difference in your skill levels, and

(b) that superficial picture is only true if you consider the very narrow perspective of a single player. If you look at the system as a whole, it's stronger when the tiebreaker system can tell not only whether you played better or worse than your opponent but by how much.

[Frydaddy 8]

I think it's becoming painfully clear that we're not actually discussing the same things here.

 

Most of us are considering the merits of X-wing's SoS-based tournament system and X-wing's MoV-based tournament system.

 

Frydaddy is at least in the ballpark of X-wing's MoV-based tournament system, if an excessively pessimistic view of it...  but I'm not convinced he's ever even read X-wing's SoS-based tournament system, because he has yet to describe anything remotely like it.

 

At six pages in, if someone hasn't bothered to educate themselves to even the most basic familiarity with the topic at hand, they're obviously uninterested in doing so.

Are you sure you understand it? Here let me spell it out.

I am 2-1 going in to the final round. and my loss was due to bad dice rolling vs a 1-2 turtling opponent in the second round( we were both undefeated going into the second round) He killed 65 points and I killed 50, diff of 15. So I get an 85 and he gets 115 and 5 points for the match win. Now going into the last round I get paired down to an "easy" opponent. I crush this guy's hopes and dream and wipe him off the table, I kill all 100 pts. He got in a few lucky shots and killed 30 points. Diff of 70, so I walk away with 170 and the match points of 5. and he walks away with 30. Now at Table 1 the two undefeated guys are going at it and it was a close game. Undefeated guyA kills 55 and undefeated guyB destroys40, which is a diff of 15. guyB and I now have the same record, 3-1. We were close in MOV all day. But because I got paired down, I took advantage of smashing a novice player and with my MOV points of170 of that game my MOV total is higher than guyB. Thank god I got paired down because 2nd place paid out $40 and 3rd place got $10.

[Buhallin 4,520]

I've said repeatedly I understand it.  I've acknowledged the potential for that to happen.  Multiple people have pointed out the rather contrived nature of the example, but even at that accept that it's still better than what SoS can do.

 

So yes, I understand it.  Unlike you, I've been paying attention for the last six pages, and since you've made not a single other point in the last six pages, I understand it so well at this point that I'm hoping it happens to you in every. single. event.  Just so I can chuckle.

[VanorDM 11,599]

Now going into the last round I get paired down to an "easy" opponent. I crush this guy's hopes and dream and wipe him off the table, I kill all 100 pts.

I'm pretty sure that there's a fairly major flaw in this example...

If you're in the final round, how likely is it that you're going to get paired against someone that you'll crush like that? I mean the whole point of Swiss is that someone with that little skill, shouldn't be in the 2-1 group in the first place.

Pretty much every example you've listed so far, seems to break the basic premise of how a Swiss Tournament actually works. How many unskilled newbies go 2-1 in the first place?

Edited July 31, 2014 by VanorDM

[Vorpal Sword 14,685]

I think it's becoming painfully clear that we're not actually discussing the same things here.

 

Most of us are considering the merits of X-wing's SoS-based tournament system and X-wing's MoV-based tournament system.

 

Frydaddy is at least in the ballpark of X-wing's MoV-based tournament system, if an excessively pessimistic view of it...  but I'm not convinced he's ever even read X-wing's SoS-based tournament system, because he has yet to describe anything remotely like it.

 

At six pages in, if someone hasn't bothered to educate themselves to even the most basic familiarity with the topic at hand, they're obviously uninterested in doing so.

Are you sure you understand it? Here let me spell it out.

I am 2-1 going in to the final round. and my loss was due to bad dice rolling vs a 1-2 turtling opponent in the second round( we were both undefeated going into the second round) He killed 65 points and I killed 50, diff of 15. So I get an 85 and he gets 115 and 5 points for the match win. Now going into the last round I get paired down to an "easy" opponent. I crush this guy's hopes and dream and wipe him off the table, I kill all 100 pts. He got in a few lucky shots and killed 30 points. Diff of 70, so I walk away with 170 and the match points of 5. and he walks away with 30. Now at Table 1 the two undefeated guys are going at it and it was a close game. Undefeated guyA kills 55 and undefeated guyB destroys40, which is a diff of 15. guyB and I now have the same record, 3-1. We were close in MOV all day. But because I got paired down, I took advantage of smashing a novice player and with my MOV points of170 of that game my MOV total is higher than guyB. Thank god I got paired down because 2nd place paid out $40 and 3rd place got $10.

But it turns out that while all four of your opponents wound up at 2-2, giving you an SOS of 40, Undefeated Guy B's opponents went 0-2 (dropped), 1-3, 2-2, and 4-0, giving him an SOS of 35. So it turns out that your noob-stomping spectacular in the Fantasyland Store Championship would still have put you ahead of Undefeated Guy B, despite the fact that his only loss was to the eventual winner.

Quelle surprise! How can this be?

It probably has something to do with the fact that in any case your poor, naive, dewy-eyed player was still good enough to be 2-1 when he met you, and for some reason you're still assuming that tabling another player in your score group is somehow always unfair to the rest of your score group. Again: what we've observed is that Dewy-Eyed Noob performed relatively well overall (2-2 isn't a record to be ashamed of, particularly if both of his losses were to players that performed well), and that you tabled him in your match. Is that because you're very good, or because he's very bad? And how do you know?

 

Pretty much every example you've listed so far, seems to break the basic premise of how a Swiss Tournament actually works.

How many people will have to point this out before he figures out that maybe he needs to rethink his position?

Edited July 31, 2014 by Vorpal Sword

[Buhallin 4,520]

 

Quelle surprise! How can this be?

 

I well and truly LOL'ed.

[VanorDM 11,599]

How many people will have to point this out before he figures out that maybe he needs to rethink his position?

3 Billion. I base that number on the fact that we're as likely to have that many people post here as he is to change his mind.

I don't know why he has so much issue with MOV vs SOS. But given the examples he comes up with, it's clear he's grasping at straws to make his point. He has an axe against MOV for some reason and logic just does not matter to him. He's made up his mind and that's that.

That or else he just has such a poor grasp on how Swiss, SOS and MOV works that no amount of explaining it will matter.

[Slugrage 5,007]

Anyone who didn't have poutine for lunch won't be able to figure it out. That's my guess.

[Buhallin 4,520]

 

I don't know why he has so much issue with MOV vs SOS. But given the examples he comes up with, it's clear he's grasping at straws to make his point. He has an axe against MOV for some reason and logic just does not matter to him. He's made up his mind and that's that.

 

I honestly think he's so wrapped around a specific bad outcome of MoV that nothing else matters.