Loss column/win column

[NeilT]

I was interested in the loss column/win column conversation in the Game Zone, in part for its acrimony, which sucked, and in part because I've usually heard it stated to the effect that the loss column is more important than the win column, not that the win column doesn't matter.  I spent some time reading through articles on the theory behind the statement, and found this, which I thought was an absolute gem:

http://www.thesandb.com/sports/assessing-the-risk-assessment-in-the-loss-column.html

Most of all, I thought it was interesting that it appeared in a college newspaper.  It's a very smart discussion.

[HudsonHawk]

I was interested in the loss column/win column conversation in the Game Zone, in part for its acrimony, which sucked, and in part because I've usually heard it stated to the effect that the loss column is more important than the win column, not that the win column doesn't matter.  I spent some time reading through articles on the theory behind the statement, and found this, which I thought was an absolute gem:

http://www.thesandb.com/sports/assessing-the-risk-assessment-in-the-loss-column.html

Most of all, I thought it was interesting that it appeared in a college newspaper.  It's a very smart discussion.

Not a bad breakdown.  It hints at one of the reasons the loss column is what matters, but never really gets there, and that's when one team has played more games, looking at the wins column automatically assumes the team trailing will play .500 in the differential games.  At the top of the standings, teams are always better than .500.

[NeilT]

Not a bad breakdown.  It hints at one of the reasons the loss column is what matters, but never really gets there, and that's when one team has played more games, looking at the wins column automatically assumes the team trailing will play .500 in the differential games.  At the top of the standings, teams are always better than .500.

I was actually thinking of it this morning as what happens at season end, if one team is a half game up.  I think what happens is that the team a half game down has to make up the lost game, and the risk of loss in the make-up game is the easiest way to see the value of the half-game lead in the standings.  If they're winning 75% of their games at the end of the season, then maybe there's only a theoretical 25%  chance of loss.  But of course that depends on whether their ace is ready to pitch, and whether they're having to make up the game against the 2014 Houston Astros or the 1927 Yankees.  It also affects who they have available to throw into a possible game against the team that's a half-game up.  They always face a risk of loss in that game.

If they're not required to make up the game, then, well, it's only the loss column in that situation that matters, and the two teams go straight to a single elimination season end.  I don't think that's what happens.

That risk of loss is easy to see in that situation, but it's also there 18 days from the end of the season. 

[HudsonHawk]

I was actually thinking of it this morning as what happens at season end, if one team is a half game up.  I think what happens is that the team a half game down has to make up the lost game, and the risk of loss in the make-up game is the easiest way to see the value of the half-game lead in the standings.  If they're winning 75% of their games at the end of the season, then maybe there's only a theoretical 25%  chance of loss.  But of course that depends on whether their ace is ready to pitch, and whether they're having to make up the game against the 2014 Houston Astros or the 1927 Yankees.  It also affects who they have available to throw into a possible game against the team that's a half-game up.  They always face a risk of loss in that game.

If they're not required to make up the game, then, well, it's only the loss column in that situation that matters, and the two teams go straight to a single elimination season end.  I don't think that's what happens.

That risk of loss is easy to see in that situation, but it's also there 18 days from the end of the season.

If two teams have played an unequal number of games, and that one game would decide who makes the postseason, it's always played.  It's only not played if both teams are eliminated or both teams have clinched a postseason spot.

[NeilT]

That's what I thought.

[BizidyDizidy]

Not a bad breakdown.  It hints at one of the reasons the loss column is what matters, but never really gets there, and that's when one team has played more games, looking at the wins column automatically assumes the team trailing will play .500 in the differential games.  At the top of the standings, teams are always better than .500.

Yes, right now the Rangers are at a .535 winning percentage. Not sure that means the loss column is the only thing that matters, but I'm no mathematician.

[HudsonHawk]

Yes, right now the Rangers are at a .535 winning percentage. Not sure that means the loss column is the only thing that matters, but I'm no mathematician.

It's what matters for following who's in the lead with games still remaining.  At the end of the day, all teams play the same number of games, so the wins and losses will have equal weight, simply by virtue of their mathematical relationship.  But when you're trying to determine "who is leading?" with 10, 20, 60 games left, what matters is how much in control are you personally, and how much opportunity to make up ground you have, if you are trailing.  If there is a difference in number of games played, that is determined by the loss column.

[NeilT]

Yes, right now the Rangers are at a .535 winning percentage. Not sure that means the loss column is the only thing that matters, but I'm no mathematician.

But that's the win percentage for the season.  For August and September they're at .642. 

[BizidyDizidy]

But if I am 70-62 and you are 69-62, I am (probabalistically) leading. I agree it's not really technically right to say I'm leading by half a game (more like .473 of a game), but I've accomplished something you've only got a 53% chance of accomplishing.

On the concept of "control your own destiny", I think the article does a good job of explaining that's really more a product of the biased way we think about these types of things.

[HudsonHawk]

But if I am 70-62 and you are 69-62, I am (probabalistically) leading. I agree it's not really technically right to say I'm leading by half a game (more like .473 of a game), but I've accomplished something you've only got a 53% chance of accomplishing.

On the concept of "control your own destiny", I think the article does a good job of explaining that's really more a product of the biased way we think about these types of things.

Of course given the same number of losses, you'd rather have more wins.  But what if you are 72-61 and I am 71-60?  Who is leading?

[NeilT]

But if I am 70-62 and you are 69-62, I am (probabalistically) leading. I agree it's not really technically right to say I'm leading by half a game (more like .473 of a game), but I've accomplished something you've only got a 53% chance of accomplishing.

On the concept of "control your own destiny", I think the article does a good job of explaining that's really more a product of the biased way we think about these types of things.

It is exactly right to say you're leading by a half game, because you are.  What you can't identify (or if you're the leading team, control) is what the risk is in that half game--there are just too many variables.  Even if you place the half game in the absolutely purest situation, a one run make-up game the day after the season ends, you still don't know who they're playing, or who's available to pitch, or  what the make-up game's going to do to their rotation, so you don't know whether it's a .473 risk or a 5% risk or a 95% risk.  What you can say is that it's a baseball truism to say that only the loss column matters because of the risk in the unplayed game.

[HudsonHawk]

It is exactly right to say you're leading by a half game, because you are.  What you can't identify (or if you're the leading team, control) is what the risk is in that half game--there are just too many variables.  Even if you place the half game in the absolutely purest situation, a one run make-up game the day after the season ends, you still don't know who they're playing, or who's available to pitch, or  what the make-up game's going to do to their rotation, so you don't know whether it's a .473 risk or a 5% risk or a 95% risk.  What you can say is that it's a baseball truism to say that only the loss column matters because of the risk in the unplayed game.

I think the truism in any sport is that if you're leading and you win out, you finish in first place.  If you win out and finish in second place, you weren't leading.

[NeilT]

Of course given the same number of losses, you'd rather have more wins.  But what if you are 72-61 and I am 71-60?  Who is leading?

No one, but that would be treated in the standings as a tie.  The marginal difference in win percentage is treated as irrelevant, and should be.

[homer]

In the last week or so of the season, the idea of 'only the loss column matters' becomes the only concern because there is some realistic possibility that you could 'win out' and 'control your destiny'.

In early September, or really any other point in the season except the last week, the question is 'what is the relative distance between two teams' and the number of losses is only part of that equation.

[HudsonHawk]

No one, but that would be treated in the standings as a tie.  The marginal difference in win percentage is treated as irrelevant, and should be.

We're not talking about "standings" as in how the teams are arranged on the page. We're talking about the philosophy of "what do I have to do with my remaining games to finish in first".  If one team is 120-41 and the other team is 2-0, technically the latter is first in the standings.

[NeilT]

I think the truism in any sport is that if you're leading and you win out, you finish in first place.  If you win out and finish in second place, you weren't leading.

But in the half-game scenario, if both teams won out in the regular season, then they would have played the same number of games, and  their wins and losses would, at the end of the day, have exactly the same value.  That's exactly what's happening when the teams face each other.  The difference is that if you're on the leading team in that scenario, the one thing you can't control (unless the teams are facing each other) is how the other team plays, and that's easy to spot in the loss column.  And, as you're right to point out, the trailing team's risk of loss is less because they're playing well, so losses really do carry more value. 

[HudsonHawk]

In the last week or so of the season, the idea of 'only the loss column matters' becomes the only concern because there is some realistic possibility that you could 'win out' and 'control your destiny'.

In early September, or really any other point in the season except the last week, the question is 'what is the relative distance between two teams' and the number of losses is only part of that equation.

Right.  No one talks about this in May, just like no one talks about magic numbers, even though they technically exist.  This is strictly to look at remaining games among teams that are close together. 

[HudsonHawk]

But in the half-game scenario, if both teams won out in the regular season, then they would have played the same number of games, and  their wins and losses would, at the end of the day, have exactly the same value. 

Again, when the season ends and all teams have played the same number of games, both wins and losses figure equally.  The exercise is how to look at it before that point.

[The Spleen]

Your opinion on the importance of the win column or the loss column may depend on your expectation for the team or teams involved. Will those unplayed games be mostly wins or losses?

Right now it appears that extra unplayed games would be more of an asset for the Rangers than the Astros.

[NeilT]

We're not talking about "standings" as in how the teams are arranged on the page. We're talking about the philosophy of "what do I have to do with my remaining games to finish in first".  If one team is 120-41 and the other team is 2-0, technically the latter is first in the standings.

I haven't really thought about it then, but that's exactly why the half game matters.  If I'm 120-41 (oh I wish), and you're 2-0, the risk is clearly on your team to play through to 120-41 in the remainder of the season.  If I'm 120-41 and you're 119-41, or 72-61 and 71-60, I'd always rather have the half-game up because you face the risk of being 119-42 or 71-61, and I don't.  That risk, whatever it may be, is the value of the half game.  I might rather have your team, because you're better than me and you're going to overtake me, but that's a different issue.

ETA:  and you're going to overtake me in both the wins and losses column.

[HudsonHawk]

I might rather have your team, because you're better than me and you're going to overtake me, but that's a different issue.

Not really, because that's what we're talking about here...can you overtake me?  What do I need to do to ensure you don't?  This only matters late in the season.