clock menu more-arrow no yes mobile

Filed under:

Indicators Pointing Up For Astros

Examining Statistical Leading Indicators Which Point To Teams' Improvement

Al Bello/Getty Images

As the off season begins, let's have some fun with leading indicators of improvement or decline in MLB teams' win-loss records for next season.  This is particularly fun because several leading indicators are pointing upward for the Astros.

Leading Indicators--An index of measurable indicators designed to predict economic activity six to nine months in advance. Leading indicators are measures which change before the economy shows a particular pattern or trend.

For the most part, I have relied upon leading indicators of next season records developed by Bill James thirty years ago.  James, perhaps the best known pioneer of sabermetrics,  observed in his baseball abstracts that teams may be poised for a breakthrough or a decline based upon the number of indicators which are pointing up or down.  A 2006 Hardball Times article discusses the indicators developed by James.  I have made a few additions and subtractions to his list of indicators, and I attempted to quantify the extent that the indicators apply to particular teams so that the results can be summed and ranked.

Like economic leading indicators, these indicators are not always accurate predictors.  After all, we don't know what kinds of personnel changes will be made by each team, whether through trade or free agent signings.  But the indicators may tell us whether the team is facing a headwind or a tailwind in its efforts to compete next season.  For example, this year's World Series champs, the Royals, have a number of negative leading indicators; but that doesn't mean they are doomed next season.  But it may mean that the Royals' management needs to make extra efforts to improve their team in order to overcome factors which point to a likely decline in the team's record.


Since I am three paragraphs into the article, I don't want to bury the lede.  So, I'll go straight to the results.  The Astros are ranked first in indicators, with +27 points, followed by the Indians, A's, Phillies, and Rays.  The five worst ranked teams in leading indicators are the Dodgers, Angels, Royals, Pirates, and Cardinals (-17 to -37 points).   The tabulation of team "points" is provided at the end of the article.  There are many ways to tabulate the points for ranking the teams, and I'm sure that one can quibble with my method.  My ranking method isn't perfect, but I think it provides a reasonable measure of the direction of indicators for each team.  Generally speaking, I utilized 10 points as the maximum positive or negative points for each indicator.

Each of the indicators will be discussed below.

Variance from Pythagorean Record

Bill James created the Pythagorean record formula (Pythag) to project teams' Win-Loss records based on the season total of runs scored and runs allowed.  Although the Pythag formula has been tweaked by analysts since then, the relatively simple method has proven to be a reliable predictor of wins and losses, particularly if only runs allowed and runs scored are known.  I utilized Baseball Prospectus' adjusted standings for this article.

Teams which substantially under- or over- perform the projected W-L record are more likely to regress in the direction of the pythag results generated in the next year.  Deviations from the Pythag record reflect random variations in the distribution of runs allowed and runs scored from game-to-game. This type of variation, to a large extent, is not likely to be repeated.  In a 2013 article analyzing teams over a ten year period with negative WPA , I found that, on average, the negative Pythag deviations regressed to within one half game of the Pythag result in the following year.

As a result, team management should not rely upon a continuation of the Pythag over- or under-performance trend.  And this means that teams which beat the Pythag may have to overcome the headwind of regression in the next season, while teams which fell short of the Pythag projection may not face the same disadvantage of run distribution in the next year.  With the exception of efforts to improve late inning bullpen performance, studies indicate that Pythag variation is largely outside the control of the teams' front offices.

I have updated this indicator by using the Baseball Prospectus second-order win-loss records, instead of just a straight Pythag.  The second order record adds projected runs allowed and runs scored to the Pythag formula, based on the teams' underlying pitching and batting performance.

According to B-Pro's second order win-loss record, the Astros' actual record fell 12 wins short of the projection (i.e., the Astros would have experienced a 98 win season without the deviation).  On the other hand, the Royals' and Twins' actual records were 7 and 10 wins higher than the projection.  Even if one assumed that the deviations are only  partially due to bad luck, clearly regression will make the Astros' task easier and the Royals' and Twins' effort to improve more difficult.

Because of the Royals' success, it has become popular to speculate that the Royals' contact hitting is responsible for the team's over-performance.  However, I found no correlation between pythag variance (as well as variance with 2d order win percent) and contact hitting (as measured by K rate).  In addition, BABIP is not correlated with the variance.   Similarly, I found minimal correlation between team variance in pythag or second order wins and home runs.   However, an inverse correlation exists between deviations to the second order win-loss percentage and teams' walk rates (R-square = -.46).  This correlation disappears for deviations from the basic pythag record.  Given the difference in correlation when projected data is added, possibly B-Pro's projections associated with the run impact of bases on balls were inaccurate in 2015. That said, I doubt that this significantly detracts from my use of B-Pro's second order standings.

For this exercise, I attributed 10 points or -10 points to the highest variances, and positive or negative 5 points to the remaining variances greater than 1 win.


This indicator is fairly straightforward.  Young players are more likely to improve than old players.  Therefore, a team which is younger, overall, has a better chance at improving.  I have used the rankings of team average age as published by ESPN.  Because this ranking is based on the average age of the 25 man roster on opening day, it isn't a perfect measure.  Some teams may become younger as they call up minor league players later in the season.  But it's the best measure I have available. I suspect that teams with notable call ups, like the Astros and Cubs, would score better on an average age measured at mid-season.

The oldest team is San Francisco, and the youngest team is Arizona.   I assigned +10 points to the six youngest teams, and +5 points to the next nine youngest teams.  Similarly, I assigned -10 points to the six oldest teams and -5 points to the remaining teams above median age.  The Astros fall in the +5 group.

AA and AAA Records

The quality of an organization's upper minor league teams bears upon the major league team's ability to add young players from year to year, as well as within the season.   Bill James originally utilized only the organization's AAA win-loss record as an indicator.  However, as noted in the previously cited Hardball Times article, based on current practice the AA record may also be relevant, since teams sometimes call up minor leaguers from the AA level.  In addition, a player may begin the season in AA and subsequently will be promoted to AAA and then the ML club during the season (see, Carlos Correa). Therefore, I split this indicator into two segments for AA and AAA, with plus or minus 5 points available for each.

I arbitrarily selected a .550 record or better as indicative of excellence and .450 or below as indicative of weakness.  A positive or negative 5 points is awarded to each organization based on excellence or weakness in the AAA or AA segment.

The Astros are the only team to score a combined 10 points for AA and AAA records.  Corpus Christi and Fresno were both among the winning-est affiliates for AA and AAA leagues, respectively.  Boston and Miami are the only teams to score a combined -10 points based on AA and AAA records.

Law of Competitive Balance

The title comes from Bill James, but this could also be called regression to mean.  Basically, regression is pushing all Win-Loss records toward .500.  As a result, it is easier for teams with weak records to improve; improvement becomes more difficult as the team's winning percent climbs. This is an admonition for winning teams: don't rest on your laurels.  Because competing teams will seek to improve, a winning team is likely to decline in the next year if it does nothing in the off-season.  I assigned points based on the team's Win %; the points reflect the percentage point difference relative to a .500 record.  A negative value applies to a winning percent, and a positive value applies to a losing percent.

The Astros receive  -3 points.  The Reds and Phillies are assigned the highest positive values, and the Cardinals are assigned the most negative value.


I added this leading indicator to the James list.  Fangraphs has a "clutch" calculation within the win value family of stats.  The measure is based on the extent that players' performance is better in high leverage situations compared to the players' overall performance.  Clutch is a measure of favorable and unfavorable sequencing of events.  By and large, the sequencing of performance events is mostly random, meaning that it is not predictive of the future.  Therefore, the advantages or disadvantages of clutchness are subject to regression, meaning that both positive and negative clutch scores are likely to regress in the direction of average clutchness in the succeeding year.  As Fangraphs has stated:

There aren’t clutch baseball teams. There are baseball teams that perform well in the clutch, but it’s not a skill; it’s just a thing that happens sometimes.

The fangraphs article examines first half and second half clutch scores, finding that there is zero predictive value in team clutch rankings.  My previous TCB article indicates that negative clutch teams, on average, regressed to mean clutchness by 40% in the succeeding year.

I divided the clutch indicator into pitching and hitting segments.  For each segment, I assigned plus or minus 5 points for clutch scores more than one standard deviation from the mean.  The Astros had one of the worst clutch hitting values in the majors in 2015, and this resulted in +5.  The Astros' pitching had a negative clutch value, but it wasn't significant enough to receive points.

Second Half Record

Another James indicator is the existence of a significant Win-Loss improvement during the second half of the season.   I admit to some skepticism of this indicator.  I understand the logic behind the indicator when the second half improvement represents a trend which may carry over into the next season.  However, I suspect that injuries are a common reason for teams' records to change over a season; and whether the injuries occur in the first or second half of the season may have minimal impact on the next season.  However, a second half improvement may be due to minor league call ups close to mid-season, which certainly can be a positive indicator for the next season.  I limited this indicator to teams with the largest deviations between the second half record and the full season record.   Therefore, teams with a 60 percentage point difference in W% between the second half and full season records received plus or minus 10 points.

The Rangers, Cubs, Blue Jays, and Phillies received +10 based on their second half improvement, and the Braves and Reds received -10 points.

Concluding Thoughts

I won't suggest that the favorable leading indicators guarantee the Astros a good record next year.  The Astros' off-season moves will be very important in determining next season's success.  Like all teams, a lot depends on how the Astros fill the production of departing players.  For example, if Colby Rasmus signs elsewhere, the Astros' front office has to figure out how to replace the team's second highest HR hitter and top Isolated Power (ISO) batter.

But the leading indicators show that the Astros will enter the off-season with a favorable environment for improvement.  In the case of the Astros, regression is a friend.  In addition, the Astros' strong minor league system and age profile provides advantages both for depth and trade ammunition.  Here is what I find encouraging: the Astros are the only playoff team among the top 16 teams in leading indicators.  The Cubs are No. 17, with zero points.  The other playoff teams are negative, with respect to these leading indicators, which isn't unusual or unexpected, given the indicators' reliance upon regression.  But it's a good spot for the Astros.

Pythag Age AA AAA Com. Bal. Clutch H Clutch P 2d half Total
HOU 10 5 5 5 -3 5

CLE 10 10

OAK 10 -5 5
PHI -5 5 5 -5 11

10 21
TBA 5 10 5

10 5 -5 8

ARI 5 10

1 5 -5
CIN 10 5
-5 11 5
-10 16
SEA 5 5


MIA 5 10 -5 -5 6

TOR 10 -10

5 10 8
BAL -5 5 5

WAS 10 -5


CHA -5 5


COL -5 5
-5 8

BOS 5 5 -5 -5 2 5 -5
5 -10

10 0
ATL -5 10

-5 -10 -1
5 -4 5 -5
SFN 5 -10




SDN -5 -5 -5

-5 4 -5

MIN -10 -5 5

TEX -10 -5

-5 10 -14
LAN -5 -5 -5 5 -7

LAA -10 -5
-5 -3
KCA -10 -10
5 -9 -5

PIT -5 -10

SLN -10 -10