Diving Deep into Tanking

"Tanking" was widely acknowledged as a problem in Major League Baseball throughout the previous Collective Bargaining Agreement (CBA; from 2016-2021). Anti-tanking measures were a focal point of negotiations on the new CBA between the MLB Players Association (MLBPA) and the league in a reaction to the explosion of analytics in front offices and their savvy business practices across the league. This has yielded important insights and surprisingly universal free-agent spending behavior. Analysts, writers, and pundits echo what current and former players have been saying more loudly in recent years: some teams are not trying to win.

The most straight-forward evidence many use to back this claim is either free-agent spending data, team payroll data estimates, or both. In support of this, team payroll roughly correlates with win percentage. This can make sense: teams that don't spend, don't win, year-to-year, on average.

That correlation has strengthened since the last CBA, though it is highly variable, year-to-year, suggesting payroll isn't necessarily a good proxy of an organization's effort to win. Including more data, Figure 1 reaches back further to when well-known "tanks" began around 2011 to reveal an even weaker correlation between all teams' payrolls and their corresponding win percentages.

Figure 1. Linear regression of MLB team win percentages with team payrolls. Coefficient of determination (R^2) = 0.176. Data from Baseball Reference.

Although a flawed conclusion, this could provide cover for blatantly cheap front offices by showing successful teams can have low payrolls and unsuccessful teams can have high payrolls. Inherent to this topic is dissecting the intent to win (impossible to prove) from actually winning. So if payroll is not a good measure of tanking via intent to win, what is?

Is there a statistic that enables us to identify tanking teams? Does it involve multiple features or other statistics? Can a statistic be developed that would enable baseball fans to recognize "rebuilding" and "tanking" processes beyond the clubs at the bottom of payroll lists? Can we use this to evaluate the rebuilds? The point of this article is to try to do that: to define and evaluate tanking.

To do this, I'll use data from 2011-2021, since Jeff Luhnow's Houston Astros tank-job and Theo Epstein's Chicago Cubs tank-job have been well-publicized and serve as seminal reference points. Both Luhnow and Epstein were hired after the 2011 regular season. All subsequent data will omit the 2020 season (for obvious reasons).

What is tanking?

It's innately difficult to prove deliberate attempts by organizations to lose or to not try to win, which is the de-facto definition of tanking. This idea is antithetical to what sports are all about (cue Herm Edwards clip). However, it has a layered purpose: to obtain and keep cheap talent for as long as possible. Thus, tanking is associated with the amateur draft. This isn't anything new: it's been advocated by fans of losing or rebuilding teams across the NBA, NHL, and NFL for years.

By enabling the worst teams to acquire the best available amateur talent, amateur drafts encourage parity and intrinsically incentivize losing. Tanking boils down to efficient spending. It's an extreme rebuild, where payroll is ruthlessly minimized, and cheap, young (amateur) talent can be hoarded all at once before that talent truly matures and balloons the payroll.

Therefore, we can establish multiple ("duh") tanking characteristics: (1) losing to an extreme degree over consecutive seasons; (2) cutting payroll before or while losing; and (3) not investing in expensive free-agent veterans. Talent acquisition strategies through international signings or the amateur draft are crucial to this process, but also to all non-tanking efforts. So they won't be included as characteristics but will be worth noting. For convenience of data acquisition, we will not consider young talent acquired via trades or the Rule 5 draft. These are crucial to rebuilding processes, but very difficult to track and quantify. So, we'll consider only the aggregate stats - total picks/signings, etc. Dollars spent on international free agents are freely available. Bonuses spent on draft picks are far harder to come by for all selections, so monetary data was omitted from this analysis.

When aiming to identify a team actively tanking, it makes sense to use the '12 Cubs and Astros- the clubs that popularized tanking- as benchmarks. Both lost 96+ games in multiple seasons (Figure 2A), and had similar payroll (Figure 2B) and free-agent spending (Figure 2C) trends.

Figure 2. Cubs and Astros' "tanking" trends, from 2011-present (excludes 2020). (A) Team win percentage, per Baseball Reference. (B) Team payroll as of the final day of the regular season, per Baseball Reference. (C) Mean Average Annual Value (AAV) of free-agent contracts, per ESPN.

The strikingly parallel trends for the 1st two criteria ("extreme" losing over consecutive seasons and cutting payroll) are revealing. A deeper dive shows both the Cubs and Astros shed ~$50 M in payroll in the first year of their rebuilds, another ~$30 M in the second, and they started adding payroll in the 3rd year. Both teams showed marked increases in payroll and in multiple free-agent contract statistics two off-seasons before their World Series championships. Most notable is the clear spike in free-agency contracts (in terms of AAV) in the off-season before winning the World Series (Cubs in '16, Astros in '17).

The different paths both front offices took to acquire long-term talent was also as interesting as the spending practices above. Figure 3A shows the Astros took a different amateur draft approach than the rest of the league (high percentage of college players drafted) and Figure 3B shows they surprisingly stayed away from international signings in the early stages of the tank (sans Yuli Gurriel).

Figure 3. Cubs and Astros' amateur player acquisition trends, from 2011-present (excludes 2020). (A) Percentage of college players selected in each amateur draft, per Baseball Almanac. (B) Total number of international amateur free-agent signings. Data from Spotrac.

The Astros focused on a mix of college bats by position, culminating in notable major leaguers, Tony Kemp (career bWAR 3.4), J.D. Davis (career bWAR 1.5), Ramon Laureano (career bWAR 9.7), Alex Bregman (career bWAR 26.1), and Myles Straw (career bWAR 4.0). Despite a focus on players more likely to be ready sooner than high school draftees, high schoolers Carlos Correa, Lance McCullers, and Kyle Tucker (who wasn't on the 2017 squad, but is important to the organizational narrative) were the arguably most impactful and enduring Astros draft picks leading up to the 2017 World Series run besides Bregman. Bregman was also the only college draftee from any Astros rebuilding draft class to be part of the MLB championship roster.

The Cubs, meanwhile, spent more via free-agency (Fig. 2C; Jason Heyward, John Lackey, Jon Lester, Jason Hammel, Ben Zobrist) and internationally (Jorge Soler, Gleyber Torres), though they also missed on "big fish", like Masahiro Tanaka. Still, Kris Bryant (USD, career 28.8 bWAR), Kyle Schwarber (Indiana; career 9.0 bWAR), Ian Happ (Cincinnati; career 6.5 bWAR), Javier Baez (high school; career 23.4 bWAR), and Albert Almora (high school; career 2.7 bWAR) were among the fruit of their amateur drafts. The Cubs acquired Anthony Rizzo and Jake Arrieta through trade early in the process, rounding out their 2016 World Series nucleus. Their build was more diverse, expensive, and less-sustainable than the Astros': the Astros improved their win percentage following their 2017 World Series win with a more stable payroll (Fig. 2); the Cubs have seemingly headed for another rebuild.

Despite these differences, discernible quantitative thresholds emerge: 2+ seasons of 96+ losses, consecutive years of > $10M cuts in payroll (about or under $100M, total), an emphasis on college amateurs, and $30M+ spent internationally (~1 major signing) and/or large spikes in free-agent contracts the year following an above-.500 season.

Together, these behaviors enable teams to re-direct funds and roster space gained from increased short and long-term payroll flexibility. These behaviors also concurrently enable teams to acquire the most robust investments as possible. In the Astros' case, the behaviors reflect an aggressive attempt to maximize the development and retention of those painfully-wrought investments. This is evident in the Astros' sustained success 4 years after their World Series win, with multiple World Series appearances since. The Lovable Losers... well... they disappointed and have not returned to the World Series.

Can tanking be quantified?

Given the above characteristics as tanking, can we identify other clubs that tanked or are tanking? In short, yes. However, each team is unique in its approach, making a sweeping generalization difficult. For example, no team focused on college amateurs in their drafts akin to the Astros. Also, relaxing a win total threshold to a more round 70 wins (92 losses) includes more "patchy" rebuilding processes where a team barely crossed the 70-win plateau during multiple heavy-losing seasons. Table 1 gives a snapshot of some of the rebuilds that meet this criteria.

Table 1. Rebuilding team performance and payroll from 2011-2021, omitting 2020. Data from Baseball Reference.

From this table, we can see that few teams gutted their payroll like the Astros in 2013, though some tried (or are trying; looking at you, Pirates!). According to ESPN, the Astros signed just two MLB free-agent contracts in the off-season Luhnow was hired (2012-2013): Carlos Pena (then 34) and Jose Veras (then 32) on one-year deals for a collective total of $4.75 M. Similarly, few teams have stripped their roster down to as low of a team bWAR as the '13 Astros.

bWAR gutting is surprising, but also a very "duh" idea when discussing efficient roster building. Figure 4A shows a perhaps forgotten, but not gone idea that team bWAR is strongly correlated to team win percentage. This validates bWAR as a descriptive statistic to evaluate a player's skill in one number and how it relates to team success.

Figure 4. The relationships between Team bWAR, Team Wins, and Team Payroll. (A) Linear regression of all teams' win percentages vs their cumulative bWAR for each season from 2011-2021 (R^2 = 0.822). (B) Team win percentage plotted against team payroll efficiency. Team payroll efficiency is quantified as the ratio of cumulative team bWAR in a season to the team's payroll (in M $US) for that season. (C) Plot as in B but on the subset of "tanking" teams from Table 1. Data from Baseball Reference.

Since tanking is ruthlessly efficient spending, deriving a simple statistic, like the ratio of team bWAR to team payroll (which I'll call WARPM for "WAR Per Million") may be key to identify the timing of when to tank or when to re-invest by beginning to add payroll again. So it's not a quantification of tanking, per se.

Figure 4B shows this stat across all teams from 2011-2021 (omitting 2020) as it relates to win percentage. Perhaps more revealing is Figure 4C: explicitly looking at W% and WARPM from the teams in Table 1. Given this stat, the Astros' behavior is notable: for the first two years with Luhnow at the helm (2012, 2013), the team lost, brutally (Table 1, Figure 2A). But the Astros were ruthlessly efficient in cutting payroll and talent, without almost any FA signings (Figure 2C). This is reflected in a substantial gain in WARPM without the Astros improving their winning percentage (Figure 4C), due primarily to the Astros' abhorrent, league-worst $18.3 M payroll.

By the end of the 2013 season, their payroll was still well below average (Figure 2B; only the Marlins had smaller), but the team began to sign free-agents that off-season (Figure 2C), focusing on veteran bullpen arms. The 2014 slide in WARPM but rise in win percentage (70 wins) may be attributed to these signings and debuts of key prospects, like George Springer.

I examined team performances of rebuilding teams with WARPMs higher than 0.4 (Figure 4C) in their subsequent seasons: the '14 Astros, '17 A's, '17 Rays, and '14 Marlins all increased their win totals by 19, 6, 12, and 15, respectively. The '15 Astros, '18 A's, and '18 Rays jumped even higher from their baselines with win totals of 103, 97, and 90, respectively. The win total differences of +52, +28, and +22 for the '13-'15 Astros, the '16-'18 A's, and for the '16-'18 Rays, respectively, are striking.

Coincidentally, the majority of the clubs with WARPM > 0.4 remain distinct from the dense cluster of other teams in future seasons (Figure 4B). If not for the well-known frugality and savvy of the A's and Rays' owners and front offices, this would suggest a launch point for sustainable success out of a tank: coordinate acquiring & growing cheap talent and cutting payroll down to a WARPM of > 0.4 to enable sustainable spending and winning. This also suggests we should keep an eye on the Marlins and Orioles...

Can we evaluate tanking?

Based on the Astros, A's, and Rays' remarkable win total jumps ~ 2 years from their last tank year ('13, '16, '16, respectively), and the aforementioned Cubs' post-World Series taper, a logical approach to evaluate tanking is to sum the win totals in the years following the last tank year and normalize by the number of subsequent years not in the tank. Disqualifying the Orioles, Marlins, and Pirates currently in a tank (Table 1), Table 2 shows the "Rebuild Grade":

Table 2. Evaluating qualified tanking rebuilds from 2011-2021. Data from Baseball Reference.

This is intuitive in a lot of ways. The Astros built sustainably by minimizing spending to invest in their developmental process, whereby their roster could be supplemented by free-agency (not the primary mode of major talent acquisition, like the Cubs) which is reflected in the Astros' year-over-year team performance. However, the Cubs also tanked successfully: they won their first World Series in more than 100 years (some would argue as the only stat relevant for measuring team success) and they also stayed competitive for multiple years, as reflected by the above grading system. The Astros just did it much, much better.

This evokes a few questions to address in the future: is tanking a permanent fixture? Only two clubs have won the World Series after having done so, yet the A's and Rays have demonstrated a quick re-set can be effective in terms of catapulting a team deep into the playoffs. So if the A's and Rays can succeed by employing similar practices as tanking teams, can anyone else replicate their strategies and make multiple World Series runs? Surely, the answer to that is yes, since the Cubs and Astros (with less restricted budgets) have shown there are multiple ways to win it all following tanking. Lastly, and more rhetorically: are multiple years of losing awfully worth 1 title and a possible dip back to obscurity?