Margin of Victory

[Vorpal Sword 14,685]

I love the discussion that getting "paired against someone with a lesser record" actually means getting paired against a "lesser player."  Even the BEST player could get paired with someone who has "lesser" skill but a cold streak on with the dice can mean defeat.  Now this player who happens to have a loss, through small fault of his own, is labelled a "lesser opponent" despite being better than you are.  Now you're saying anyone who has to face this guy is "lucky" because he happened to get unlucky early on.

Just want to point out that I'm not making this assumption--adjacent score groups are fairly similar in skill level, and a match between disparate skill levels can easily happen when the better player is paired up to the next score group (particularly early in a tournament).

[Introverdant 707]

 

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.

[Vorpal Sword 14,685]

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.

The median size was something like 30 players, in which case it's likely that (after 5 rounds) there will be one or two 3-2 players who make it into the top 8.

[Introverdant 707]

 

 

 

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.

 

The median size was something like 30 players, in which case it's likely that (after 5 rounds) there will be one or two 3-2 players who make it into the top 8.

 

 

Huhm.

Okay then!

[ScottieATF 2,864]

Double Elimination would hurt participation.

Citation needed.

One of the selling points of Swiss is that even the losing players get to keep playing.

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.

I suppose it's irrelevant. If it were up to me, it'd be double elim - but Swiss is what we have so *meh*.

The majority of people playing in a Regional, Store Championship, Nationals, or Worlds have no real shot at winning or even making a cut. Most know it. They are there to play for fun, it gives them a chance to get alot of games against people they don't normally play. If you can only gaurentee a player that is just there to play, two games, they are less likely to take the time and make the trip. Nothing I'm talking about has anything to do with those hoping to make a top cut or win an event. Swiss allows players there to just get games in to keep playing even once they are out of contention. This helps general participation.

Edited July 30, 2014 by ScottieATF

[Buhallin 4,520]

 

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

 

So only the people who are going to make the cut matter?

 

There may be an argument for double elimination at large events like GenCon or Worlds where eliminated players may want to go do other things.  As a general approach for all tournaments - which are as much about building community as competition - it would be a disaster.  Making half your players have nothing to do but twiddle their thumbs after the first game or two is not a recipe for involved players.  It's also a HORRIBLE approach to building community, because those probably-new players are exactly the ones you want to encourage to come back.  Most players I know will happily take losing games to being eliminated and not getting to play.

 

Double elimination would also, in most cases, result in longer events than Swiss.  You can reasonably do 16 players in 3 rounds, reliably in 4.  Double elimination generally would take 6 to cover the same.

[Introverdant 707]

Thinking about Regionals, yeah...

there were a lot of people that had driven hours to participate.

Probably wouldn't have made the trip just to play two or three games. 

 

Okay. I'm sold. 

[doji 163]

Personal I like Sos better.

The only big difference between mov and Sos is that it no longer hurts you if people you played dropped out or if you got stuck with an early bye or you played someone with a bye.

I agree that this system punishes even skilled players due to the fact that you should only win bye 1 or 2 ships so your average mov score per game would be less then that of a bye. Not guaranteed just very probable.

[Vorpal Sword 14,685]

Personal I like Sos better.

The only big difference between mov and Sos is that it no longer hurts you if people you played dropped out or if you got stuck with an early bye or you played someone with a bye.

And it also no longer hurts you to have an opponent with a modified win, or to be matched up in the first round with a very weak player, and I could keep going but instead I'll just refer you to all of my other posts here.

I agree that this system punishes even skilled players due to the fact that you should only win bye 1 or 2 ships so your average mov score per game would be less then that of a bye. Not guaranteed just very probable.

I honestly am not sure what you mean here.

[Buhallin 4,520]

 

I agree that this system punishes even skilled players due to the fact that you should only win bye 1 or 2 ships so your average mov score per game would be less then that of a bye. Not guaranteed just very probable.

I honestly am not sure what you mean here.

 

I think he's saying that a player who gets a bye will get the full points, whereas even a good player won't, so the good player gets less MOV than the bye player.

 

Which is true, but misses that (1) this already happens with SoS byes, and (2) this is actually a GOOD thing, since part of the point of bye scoring is to jump the low player up enough that they don't just stay at the bottom.

[Buhallin 4,520]

Thinking about Regionals, yeah...

there were a lot of people that had driven hours to participate.

Probably wouldn't have made the trip just to play two or three games. 

 

Okay. I'm sold. 

What?  Seriously...  WHAT?!?!

 

Dude...  you just let someone persuade you away from your original view.  Come on...  this is the INTERNET.  We don't do that sort of thing around here!

[WonderWAAAGH 7,153]

Thinking about Regionals, yeah...

there were a lot of people that had driven hours to participate.

Probably wouldn't have made the trip just to play two or three games. 

 

Okay. I'm sold.

He can be taught!

[Introverdant 707]

 

Thinking about Regionals, yeah...

there were a lot of people that had driven hours to participate.

Probably wouldn't have made the trip just to play two or three games. 

 

Okay. I'm sold. 

What?  Seriously...  WHAT?!?!

 

Dude...  you just let someone persuade you away from your original view.  Come on...  this is the INTERNET.  We don't do that sort of thing around here!

 

 

I still think Strength of Schedule and Margin of Victory are super weird and contrived. 

 

Would prefer to see top 8 go to elimination style brackets, or something. 

I suppose that would mean a 2nd day in most cases. Hmmmmm...

 

There's really no perfect solution here, is there?

Edited July 30, 2014 by Introverdant

[StevenO 2,996]

 

I love the discussion that getting "paired against someone with a lesser record" actually means getting paired against a "lesser player."  Even the BEST player could get paired with someone who has "lesser" skill but a cold streak on with the dice can mean defeat.  Now this player who happens to have a loss, through small fault of his own, is labelled a "lesser opponent" despite being better than you are.  Now you're saying anyone who has to face this guy is "lucky" because he happened to get unlucky early on.

Just want to point out that I'm not making this assumption--adjacent score groups are fairly similar in skill level, and a match between disparate skill levels can easily happen when the better player is paired up to the next score group (particularly early in a tournament).

 

 

Comment was left after the first page but early pairing can have a lasting effect no matter what happens afterwards or what scoring system is used.

 

I may not be sold that MOV is always better than SoS but I do believe it give a better feel for the "quality" of wins.  If FFG figures SoS off of actual W/L records that would be one thing but when its only based on "wins" it gets off too easily.  In almost any case bade games can happen either through random luck with the dice or even just because of unlucky pairing.

[iPeregrine 685]

And it also no longer hurts you to have an opponent with a modified win, or to be matched up in the first round with a very weak player, and I could keep going but instead I'll just refer you to all of my other posts here.

On the other hand now it helps you to be matched up with a weak player in the first round, so the first round random pairings still have a major impact. A player who gets randomly matched with a good opponent in the first round will probably have a fairly small margin of victory, while a player who randomly gets matched up with a weak player will have a huge margin of victory. In later rounds the difference in margin of victory becomes less significant because the system matches players of roughly equal skill and/or list strength, but that just means that the first round has a much bigger impact than any other round.

There was a similar problem with this in 40k tournaments back when most events didn't run enough rounds to pick a clear winner and used margin of victory to break the tie between the undefeated players. The most important part of winning the tournament wasn't competing in the final rounds, it was getting lucky enough to have a baby seal clubbing for your first match and generating such a huge margin of victory that nobody could catch up.

[Frydaddy 8]

At our Imdaar alpha tourney we had:

1 undeafted

And 4 guys with a record of 3-1(or maybe it was 4-1, but it doesn't matter-4 guys each with one loss)

So when we broke to the top 4, one of the one loss guys wasn't going to make it. And the guy who didn't make it had his loss against a player who was not in the top 4. His loss was against a player who went on to lose. The other 3 guys w one loss each all lost against one of the other top 4 players. When we broke to the top four, the leftout 4-1 player was wondering why he didn't make it, with strength of schedule it was very easy to answer this player, "You lost to someone that had a worse record than you." And with that response the talking under the breath of "I went 4-1. I should be in the top 4." And all the other crying Ended.

Now with MOV this will not be the case. I don't know what to tell them, except, "it's easier to calculate."

[Mace Windu 1,101]

At our Imdaar alpha tourney we had:

1 undeafted

And 4 guys with a record of 3-1(or maybe it was 4-1, but it doesn't matter-4 guys each with one loss)

So when we broke to the top 4, one of the one loss guys wasn't going to make it. And the guy who didn't make it had his loss against a player who was not in the top 4. His loss was against a player who went on to lose. The other 3 guys w one loss each all lost against one of the other top 4 players. When we broke to the top four, the leftout 4-1 player was wondering why he didn't make it, with strength of schedule it was very easy to answer this player, "You lost to someone that had a worse record than you." And with that response the talking under the breath of "I went 4-1. I should be in the top 4." And all the other crying Ended.

Now with MOV this will not be the case. I don't know what to tell them, except, "it's easier to calculate."

 

So basically in this example:

 

SOS says through no fault of their own the player who missed out did so because of random pairings and SOS and no matter the fact he played flawlessly he missed out anyways because SOS pairings says so;

 

OR

 

under MOV you tell the guy that he didn't make it because he didn't destroy as many points as the guys that did make it and he lost more points than the guys that did make it.

 

pretty sure MOV allows you to dictate your own destiny far more than SOS ever can.

 

Thats the way I read it, though someone here will likely tell me Im doing something wrong.

[Gundog8324 440]

 

So how would you calculate an SoS where one of the players drop?

If an opponent drops, his score ends… how is it different than all his remaining matches are losses?

"How is it different than all his remaining matches are losses?" Really? Hmmm lets see. A player goes 1&3 and then drops. This player has a 33.33% win percentage. If he keeps playing and loses again his win percentage is 25%. I'd rather have an opponent with a 33.33% than a 25%.

 

Hmm quick question since when does 1 win out of 4 games played give you a 33% win percentage? 1/4 is .25 not .33 and 1/5 is .2 not .25

 

That said under either system there will be people who will feel like they were screwed and either you are penalized for playing a "weaker" opponent or rewarded.

 

Frankly I like MoV better because it feels closer to a "modified loss" you don't get the point for the tournament standings but keeping it close should be worth something.

[MajorJuggler 7,752]

This is just semantics, but calling it "margin of victory" is somewhat misleading. It is actually:

 

(Points For) - (Points Against) + 100*( # of rounds)

 

The +100 points each round is completely unnecessary. The only reason to do that is to make everything a positive number. Maybe they are afraid that people don't know how to add negative numbers?

 

Recording the PF and PA, and then calculating the difference is exactly the same thing.

 

Taking out the "free" 100 points per player per round, the total number of MoV points in the player pool must sum to zero, at least until byes are considered. Get a bye and it treats you as having won by 50 points, but there is no corresponding MoV loss for another player. Point creation out of thin air!

 

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.

[WonderWAAAGH 7,153]

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?

Edited July 31, 2014 by WonderWAAAGH

[ScottieATF 2,864]

Other FFG games have two categories of Byes.

Awarded and Odd Number.

Awarded grants full SoS possible, Odd Number grants 0. For instance.

Hence them being called Super Byes by some

[That One Guy 2,242]

I like the new system. It rewards players who manage to dominate the match, while at the same time rewarding losing players who fought tooth and nail. This way even losing in a very close match can still be almost akin to a win later in the tourney.

 

The other interesting implication is in heavily loaded ships. On the one hand, if they got focused down then you lose a lot of points. But if you manage to keep them alive then they can be really worth it! For instance, in my last tournament the match I came closest to losing saw 2 of my 3 ships killed. However, the one that survived was Rexler Brath, a 47 point ship.

[Spike IT 57]

One thing that could be done to mitigate the weirdness of the randomness would be to seed players in each round based on tiebreaker.  FFG's system doesn't actually do this for some reason.  You group players only by tournament points, then pair them randomly within each group.

 

If you seed by tiebreaker, skew from the random pairings tend to work themselves out.  If you demolish that random newbie, you end up high on the list and face other big winners.  If you squeak by because you had a harder opponent, you end up lower and get opponents who also didn't do as well in the first round.

 

I've never understood why they don't seed by tiebreaks, either here or SoS.  For SoS it maybe makes it easier than having to calculate the full SoS each round, but differential is far easier to track.

 

It's actually worth noting that you CAN use the tiebreaks.  The tournament rules specify that players will be matched to others with the same number of tournament points, but it doesn't go into detail on how to do that...  so you can use the tiebreak and be perfectly within the rules.

 

This is what chess does with their rating system.  Players are grouped based on similar scores, but then their rating determines who they get paired against in the pairing group.  Pairings start with the highest rated player and work down until the middle is reached, then start again for determining the opponent.  So the top rated player in the scoring group plays the "top of the bottom", 2nd highest rated plays the "2nd of the bottom", etc.  If there is an odd number of players, the player that is left of is the lowest rated player - and they will get paired against the top rated player in the next lowest scoring group.

 

I have run a number of chess tournaments in the past and believe this is a fair way of running events.

 

Using current round MOV as the "rating", I think this would work very well.  It does not eliminate the randomness of round 1 when everyone has the same score and MOV, but it works starting in round 2 - and think it works better the further the tournament progresses.  I could see this being the next evolution of FFGs tiebreaker and pairing system.  Using tournament software, this is actually pretty easy to do.  It's a little more cumbersome doing it on paper, but for smaller events - it can still be done pretty easily.. 

[bothanspy 12]

The pairing method is not really important, as long as most of players play with someone of the same score group, and that no one plays against the same opponent.

 

The methods you can find are just convenience method to generate pairings by hand. If you're using an advanced software, it can scan through all the possibilities and find the best pairing possible.

 

The randomness of pairings is something that will always be present in a swiss system. Whatever the tiebreak is, it will influence the final ranking. The question is, given a player's schedule, destroyed and lost points, how do you compare it to another player with the same score?

[Spike IT 57]

In a swiss system, most players are playing someone in the same score group - that is the definition of a swiss system.

 

The pairing method is important IMO - random is not good, but is present in the swiss system as you said.  Whether you are using software or not, there should still be rules that need to be followed for the pairings to minimize that randomness - which X-Wing currently does not have. Using the current round MOV to determine rating seems like a good way to me - at least after round 1 when everyone does not have the same score and rating.

 

Yes tiebreaks will influence the final ranking.  And how you compare a players schedule to another players with the same score is the eternal question.  FFG has given us the new MOV method of determining tiebreaks, which I think is better than SoS for many of the reasons previously given.  Whatever method of tiebreaks is used though, I believe using those tiebreaks to help determine the pairings in a round helps eliminate some of the randomness that happens for games that do not have a rating system of its own.