THE BOOK--Playing The Percentages In Baseball

I dunno.  I mean, despite the vast improvement projects, I think it still qualifies as a shock to expectation (based merely on history) that a team that’s finished 28 or more games out of first every year for the past seven years might now win the division.  Has that ever happened? 

I imagine most of the teams who have gone “worst to first” throughout history were bad for a relatively short period of time.  Detroit has a couple of middling years winning 71 and 72 games before jumping near the top of the division.  Judging from B-Ref, the Rays are Pythag’d to hit 94 wins this year.  That’s almost a 30 win jump from their past three years.  I have to think that’s unique enough to account for a certain degree of “shock”, even among those who “expected” such things.

Was that really a clear consensus among “smart” forecasters?  PECOTA had them at 90, CHONE had them at 89, THT had them at 81, Zips had them at 79.  I have no idea about any other forecasters.  That averages out to close to 85, but there’s a huge swing between forecasts there. Some had the Rays as a contending team (though probably still in 3rd), and others had them well out of the race.

Clearly anyone that thought they were going to continue to be a 66 win team was ridiculous and didn’t take into account the amount of (young) talent that they had, but they’re on pace to win 30 more than they did last year.  Was it ridiculous to think they’d top out around 80 wins this season?

I am a little biased because I had them at 86 wins. But yes, it IS ridiculous to think that they would “top out” (whatever that means - team’s don’t “top out") at 80 wins, if you are a credible forecaster or if you familiarized yourself with the credible forecasters’ forecasts.

I don’t care if one forecaster had them at 10 wins and another at 160.  If the average of ALL the credible forecasters had them at 85 wins, then winning 90 games is NOT a surprise, let alone a shock.

#1, I already said that the “story” is a team that goes from last to first (although personally, I don’t consider it a “story” - big deal, a team gets better because it has a lot of young talent).

The “story” is not that a team that everyone expected to win 67 games wins 90 games (or whatever they end up winning).  Anyone who thought that they were a 67 win team knows nothing about evaluating a baseball team’s talent.

As I said, it may be a “story” for lot’s of people, but the “rest of the story” is that you would have had to be ignorant to think that they were a bad team.

I suppose it is indeed a “shock” to all the people who thought that they were supposed to win 70 games this year (which are a lot of people, including those who write about baseball for a living), but that is just a testament to their ignorance.  That is all I an trying to say.

Top out was a poor phrasing on my part.  I didn’t intend to say that 80 was their reasonable max.  But I don’t think that 80 wins would have been a terrible forecast for them coming into the season, though you do, and of course you would have won that bet. But given an 80 win true talent team, 95 wins would be surprising to me though certainly not shocking to the extent I’m sure Murray Chass portrayed it as, but that’s what you get when you read something from an “esteemed” writer such as him.

Sure, if they were projected as an 80-win team, then their current record would be a surprise (but not a shock).  Even as a true 80-win team, being 65/44 (as of yesterday) will happen around 3% of the time, not too “shocking” in my book.  About the same as a home team losing a game with a 3-run lead in the 9th (I think).  That happens all the time, and people don’t write articles about it!

And trust me, while I may have been “right” about the Rays, I completely tanked on the Padres, Indians, and Angels!

Chass is not in a one-man club.  Ask any MS writer or commentator what they thought the Rays were going to do this year, and I guarantee that 99% of them will give you a number less than 80 (if they were being honest).

To be fair to them, should we expect them to know how to do projections for players and then for teams, etc?  Of course not.  I would, however, expect them (if they were not lazy and ignorant) to apprise themselves of at least a couple of the credible forecasts out there and at least mention them in these articles (such as, “While most people did not expect the Rays to contend, there were some analysts who were a bit more optimistic...” or something like that.  Rather than, ”Everyone thought....”

I suppose it is indeed a “shock” to all the people who thought that they were supposed to win 70 games this year (which are a lot of people, including those who write about baseball for a living), but that is just a testament to their ignorance.  That is all I an trying to say.

And Vegas, since that over 72 bet I made is looking pretty good…

It is possible that there may be an inefficiency in the betting market when it comes to new (young) talent coming on board, that the serious forecasters (who had the DRays at around 85 wins) could easily trump Vegas.

Just a WAG at the moment, and an idea for someone to explore…

Even as a true 80-win team, being 65/44 (as of yesterday) will happen around 3% of the time, not too “shocking” in my book.  About the same as a home team losing a game with a 3-run lead in the 9th (I think).  That happens all the time, and people don’t write articles about it!

Love that line!

First of all,
A team improving by 20+ games, to surpass NY & BOS will be a big story.

Okay, the park factor at TB has been right around 1 the last few years, so I wont park adjust.
Their offense has scored 4.49 runs a game, below league average of 4.70 - Lets’s assume that we project their runs scored correct for the year.

Their pitching is 443 runs - 4.03 per game. 92.3% of the runs are earned. If you take each pitcher’s improvement in ERA from 2007 to 2008 and multiply it by 2008 ip you get 126 runs - if
you divide by .923 you get 137 runs. Add 137 to 443 you get 580 runs -

That is pythageros (sp?/actually Greek translation?) of 47 wins .

If you pegged their offense perfectly, and thought that the 14th best ERA team in the league would be worse than average the following year, then you have TB at less than 81 wins. I dont think that would make anyone “stupid” or “dumb”.

I dont think the guys setting the Vegas’s lines are stupid or dumb.

As we all know, the biggest turnaround from last year to this is their defense (and thus, the ERA or RA) which bad incredibly bad and incredibly unlucky last year (over 100 runs bad in team UZR).

Honestly, I don’t mean to say that anyone is ignorant or stupid for not predicting them at 80 some odd wins.  I made that clear in my last post.  No one BUT a serious forecaster who uses offensive, defensive, and pitching projections would have been able to do that.

What I meant was that these writers, after the fact at least, should have familiarized themselves with what the serious forecasters put out and then mentioned that in their “shocking story” articles.

Again, write something like, “It should be noted that not everyone thought the Rays were going to be a bad team this year.  In fact...”

And yes, Tango, the Vegas betting market for team win/losses is much more inefficient than the individual games, for various reasons.  One reason is that you can’t bet much and the vig is very high and as long as their total handle is reasonably balanced (if they put up a really bad line on an individual game, they would NOT get balanced action), they don’t care what the line is.

MGL: can you do us a favor?  Maybe this would be a great article.

At game #1 (opening day), the prior for each team is probably the Vegas line, correct?  That is, you use the forecast for each team’s W/L for the season, look at the starting pitchers, and you get the line for game 1.

At game #2, you have the prior (the W/L season forecast), the performance at game #1 (and whatever “feelings” bettors have) and you have the line for game #2.

And on and on for each game as the season unfolds.

When it comes to the Rays, did the “implied” win% for each game follow a standard “in-season Marcel” expectation, such that they give the pre-season forecast, and the in-season performances the typical due weight?

Or, did Vegas very quickly backtrack on their preseason forecast, and realize how wrong they were (i.e., took MGL, chone, pecota seriously), and make a quick adjustment?

I looked for PreSeason picks - ESPN had http://sports.espn.go.com/mlb/preview08/news/story?id=3320198 - Toronto at 85 wins, TB at 73 - Since Vegas had 72, let’s call the masses TB at 73.

HT and BP have the Rays rated very even with the Jays for the season- HT staff consensus has TB 4th, BP 3rd- Between the 2 18 picked TB better, 17 Toronto- and I dont think ANY picked them better than NY or BOS. no wins attached, but I would think that would be right around 81 wins.

http://www.hardballtimes.com/main/article/2008-tht-staff-predictions/

http://www.baseballprospectus.com/article.php?articleid=7307

Right now, they are on pace to win 97 - 35 informed experts all picked them to finish behind BOS and NY, and right now they are ahead of both of them.

The experts gave them 81, better than the dummies at 73, but still way shy of 97. It’s kind of like getting a higher F than someone else on a test.

What’s more annoying is how all the “analysts” began to question the team after their little 7 game losing streak, and began to blame all the usual suspects when dealing with a young team.

Now, the “analysts” say, “they managed to overcome their skid and have shown experience blah blah blah.”

I mean, c’mon.

And Vegas would also have to be labelled “stupid” too.  I got the Rays at 150:1 to win the World Series the opening weekend of the season at The Stratosphere in Vegas.  That ticket is currently sitting in a safe deposit box in the bank.
vr, Xeifrank

Vegas was only “stupid” on the Rays (and Seattle, BTW) because most people were “stupid” on them, and the pre-season Vegas lines largely reflect what most people with a little correction by “smart” bettors (anyone who went online and looked at BP, THT, etc. to make some of these pre-season bets.

Tango, I can guarantee without looking at the lines, that Vegas quickly caught up with the Rays’ talent.

dq, the “on pace thing” makes no sense.  After one game won, is a team “on pace” to win 162 games so anyone who projected them as a 100 win team is an idiot?  No!

The Rays are on pace to win whatever their wl% is now regressed to whatever mean you want.  I have them currently as a .520 team which means they are “on pace” to win 93 games.

Again, I am not and have never said that these “shocking story” writers or anyone for that matter, are stupid for not “knowing” that the Rays were a good team going into the season.  They are stupid, ignorant, or lazy, for now knowing that the people who actually spend more than 30 seconds thinking about a team’s projected wins for the season thought they were at least an 80, if not an 85, win team.  And THAT should be mentioned in these types of articles.

Saying that “everyone” said or thought something is NOT true at all. And even if “everyone” means “almost everyone” there is a bog difference between everyone that knows how to form an opinion on something and everyone who doesn’t, if that makes any sense.

Again, that would be like saying, wow, it is a shock that Jeter has a bad UZR this year, because everyone thought that he was a Gold Glove defender.  Again, I guess it is a shock to THOSE people, but if you are going to write a responsible article on something like this, you have a responsibility to tell the whole story, which is simply that anyone who takes doing team projection seriously thought that the Rays would be a good team this year. But that would not be such a good story I guess.

Im only quoting Vegas (and ESPN) to determine what the general public thought.

Im confused - you had them at the beginning of the year as .530 (86 wins), and now you have them as .520 - after playing .600 ball for 2/3 of a season.

How does playing .600 ball make a team worse?

A team that has won 3 games out of 5 is at a pace to win 97 games - they may not maintain that pace, but that is the rate at which they are performing. A team that has won its 1st game is on pace to win 162.

Yes, I know that “on pace” means to just extrapolate the current record, but I was making a point that you don’t want to compare, say an 88 win projection, to a team that is “on pace” to win 97 games and then conclude that it was a bad projection any more than you would take a 1-0 team and conclude that a 100 win projection for them is a terrible projection because they are on pace to win 162 games.  So if you can’t do that after one game, when is the magical point at which you can use an “on pace” win/loss to conclude whether your projection was good or bad?  There is none!  You have to take a team’s record so far and then add in the rest of the season using some OTHER win percentage. 

For example, if a team goes 1-0 after the first game, I KNOW that an 82 win projection is so far better than an 80 win projection.  However, I don’t know whether an 82 win projection is better than an 81 win or an 83 win one.  There is a mathematical answer, but I don’t know what it is. Andy, or some of the other guys can answer that.

I can guarantee you, though, that a 90 win projection for the Rays right now is BETTER than a 103 win projection even though the 103 is closer to their “pace” than 90 is.  Do you know one way I can “prove” that?  I can ask anyone here to take one or the other in an even money wager, and if they value money, they will ALL take the 90 wins and not the 103!  So therefore, using an on-pace number at ANY point in the season is NOT the best number to use to evaluate a projection.  That is the only point I am trying to make. The simple reason is that any team (that we don’t know anything about other than their w/l record) has gotten lucky (on the average) if they are “on pace” to win more than 50% of their games and vice versa if they are on pace to win less than 50%.

Why did I have them at .530 before and .520 now that they are playing at a .595 pace (or whatever it is)?  Because I pay NO attention to how they are playing!  My best estimate of how a team is going to play for the remainder of the season is gotten by projecting each player on the team and adding them all up.  If you want to throw in a “tiny” fraction of their pace (like maybe 10% record, 90% player projections) just in case there is something I am missing with the player projections, I won’t quarrel with that.  I don’t do that though.  Using a team’s in-season pace after even 120 games to project their remaining pace is like using a player’s RBI and runs scored to project their OPS. OK, bad example, but by that I mean that there is a tremendous disconnect, in the short run, between one and the other because of all the noise associated with team W/L records in even 100+ games.

Only in baseball do we make no sense of numbers.

If in a 100m race someone after 0.5 seconds is leading 6M to 3M, the leader is not “on pace” to win the race by 50M.  No one will “prorate” that.

If Federer wins the first game of the first set, he is not “on pace” to win 6-0, 6-0, 6-0.

A team that wins 3 of 5 is not “performing at a rate” of anything.  Why not say they are “performing at a differential” to end the season with one more win than loss?

Just because you can divide numbers with the racers, tennis players and teams doesn’t mean we should.

Even if the writer picked TB 3rd, like 1/2 of the BP/THT people did, their being in 1st place over NY and BOS is a big story.

Pace is the rate of activity,progress,performance.
Pace is used all the time in/during races.

A miler running the 1st lap of a mile in 60 seconds is on pace to run a 4:00 mile. (~342 1500)

A swimmer who swims the 1st 100 meters of a 400 meter race in 56 seconds is on pace to swim a 3:44 400.

Racecar driver, horses,etc…

Because the difference in your 100 meter race is the start, it is not a good example of pace. - You have 2 separate activities, the start, and the acceleration. When you are accelerating, you dont have a relatively constant pace.

Once you get past the start, pace is used in racing all the time.

A team that wins 97 of 162 games won games at a pace of .600.

I’m not saying that pace is the best estimate of how a team will perform the rest of the year. But the people who are observing TB’s performance so far are observing a .600 team, that even “good” forecasters said would be .500, and the masses said .450.

I agree, using their pace number is not the best way to evaluate a projection done before the season starts. But a sportswriter is most likely going to talk about TB’s performance so far, and how it is better/so much better than anyone first thought.

I dont think you can “guarantee” that a 90 win projection is better than a 103 for TB. I believe it to be true, and your chances of being correct are probably over 90%, and no one will take an even money wager on it. - But how about $100,000 versus a $1 on a bet?

For Federer, I dont know. At some point if he wins every game 4-0 his chances of winning 6-0,6-0,6-0 increase greatly.

My point is that no one talks about “pace” like that.  No one says that after Federer goes 6-0 in the 1st set against Nadal, that he’s on pace to go 12-0 the rest of the match.  No one says that the Jets are on pace to finish 16-0 after a 2-0 start.  But, in baseball, you get that b.s. all the time, like someone has 40 RBI after 33 games and “he’s on pace” to beat the RBI record.  No he’s not.

I just found a dollar a second ago… look, I’m “on pace” to earn 3600$ in the next 60 minutes.  Yes, if by “pace” you mean continuing my enormous streak of good luck.

Extrapolating good luck into the future is not something one should do, and it drives me bananas every time it’s done.

Help me out with the .520 versus .530? If you pay no attention to how they are playing, how do they get worse?

After 110 games, shouldnt you put some weight in the current performance in determing how good the team is? Im not saying use the .600 - Im saying I would think your assessment of the team should include some factor for current performance.

(Especially for young players)

I read the article and got the sense that Murray Chass was extremely impressed with the Rays but I don’t see much shock on Chass’ part. The Rays made some terrific changes and are now a very good team. I don’t see anything there that suggests they are playing over their heads or getting lucky.

I understand that it’s a theme of the blogosphere that veteran baseball writers are idiots, but most of what I remember reading this spring said the Rays were a real team now. Is there really that much evidence to support this statement by MGL? I haven’t looked and don’t know—I’m just wondering.

“Veteran baseball writers thought that the Rays were a bad team going into the season, because they have no idea how to evaluate a team when that team has significantly changed its personnel.  They are lazy and stupid and all they can do is look at last year’s record and make some really weak adjustments in their head.  Now that the Rays are doing well, as they were supposed to, we can see how really dumb baseball writers are.”

And when a team that has finished last every year is in first place on Aug. 1 ahead of the sport’s two dominant franchises, I’d say that’s a story.

I was exaggerating quite a bit with my original post.  Also, I did not read the Chass story.

I only wish that one of these mainstream writers would mention that most of the serious forecasters expected the Rays to do very well.  That’s all.

dq I use the current season stats of the players to update the player projections, and then I add up the player projections.  I don’t “care” about a team’s w/l record but I do “care” about how the players have played.

As to why I had them as a .530 and now a .520 team, I don’t know off the top of my head.  I must have used a more optimistic estimate of playing time for players before the season started or they had some injuries, or something like that.

To be honest, I don’t update the team win percentage very often, so they could be a .530 team going forward.  I don’t know.  That would only give them an extra half win or so.

I agree with Tango a million percent.  Of course you can legitimately argue that the technical definition of “pace” is simply to extrapolate a certain “ratio” or “rate” at any point in time, but using it for any statements or conclusions about the future is just not useful at all.

The sportscasters will casually say that so-and-so is on a pace to hit 50 HR say, after 74 games.  They are more than implying that that is a reasonable number to expect.  I also chuckle to myself and think, at what point is it OK to casually mention that.  A player hits 2 HR on the first day.  Do you casually say, “Boy, he is showing a lot of power this year.  He is on pace to hit 324 HR, but he probably won’t do that well.”

Just one of those things (among many) that bugs the hell out of me, and apparently Tango too. No big deal though.

TB has played at a .667 pace/rate/whatever since this was started - 66-44 versus 80-51.
This is without their top hitter missing most of the games.

They need to go 17-14 to get to 97 wins. If they are a .530 team they will fall one game short.