FANALYTICS: Mayberry 4.0 and Portfolio3

An unabridged version of this article appears in the 2015 Baseball Forecaster.

There was a time when 30/100/.300 meant something. Those were the benchmarks for stardom; a point of reference around which we could evaluate performance. Thirty home runs, 100 RBIs and a .300 batting average. When PEDs were peaking 15 years ago, at least two dozen players were reaching those numbers every year.

And now? Last year, the only players to accomplish 30/100/.300 were Jose Abreu and Victor Martinez.

There was also a time when a sub-2.00 ERA was virtually unachievable. Thirty-one pitchers (min. 30 IP) accomplished that feat in 2014.

A strikeout-to-walk ratio of 2.0 used to be the LIMA Plan's delimiter of success. Today, pitchers with that ratio are in the bottom 25%. Starting pitchers (min. 120 IP) with that ratio are in the bottom 15%!

So, what’s “good” any more?

I don’t know; certainly, the old benchmarks are of no help. I suppose we could arbitrarily set new benchmarks based on current conditions—30/100/.300 becomes 25/90/.290—but how long would they remain relevant?

Over the years, we’ve presented several of our metrics normalized to a league average level. PX (linear weighted power index) and Spd (statistically scouted speed) have always been presented as an index with 100 representing league average. Batters with a 120 PX, for instance, were producing power 20% above league average; batters with a 65 PX were 35% below league average.

This was intended to provide a better measure of “good” for those skills that had a tendency to shift over time. And 100 could bounce off the walls for all we cared; it didn’t matter because everyone was benched to that league average.

Now, nearly everything is shifting. The use of indices to normalize metrics to a league average is more important that ever.

Given all of this, it has become time for the Mayberry Method to acquiesce to the changing times and transition all of its components to normalized metrics.

Mayberry Method 4.0

The foundation of the Mayberry Method (MM) is the assertion that we really can’t project player performance with the level of precision that advanced metrics and modeling systems would like us to believe.

MM is named after the fictional TV village (above) where life was simpler. MM evaluates skill by embracing the imprecision of the forecasting process and projecting performance in broad strokes rather than with hard statistics.

MM reduces every player to a 7-character code. The format of the code is 5555 AAA, where the first four numerics describe elements of a player's skill on a scale of 0 to 5. The three alpha characters are our reliability grades (Health, Experience and Consistency) on the standard A-to-F scale. The skills numerics are forward-looking; the alpha characters grade reliability based on past history.

MM has gone through several iterations over the years, adjusting some of the factors. We are now at version 4.0.


The first character in the MM code measures a batter's power skills. It is assigned using the following table:

	PX            	MM   	
	0 - 49        	0    	
	50 - 79       	1    	
	80 - 99       	2    	
	100 - 119  	3    	
	120 - 159  	4    	
	160+          	5    

PX is already a normalized metric; no change is needed.

The second character measures a batter's speed skills. RSpd takes our Statistically Scouted Speed metric (Spd) and adds the elements of opportunity and success rate, to construct the formula of RSpd = Spd x (SBO + SB%).

	RSpd            MM
	0 - 39        	0    	
	40 - 59       	1    	
	60 - 79       	2    	
	80 - 99  	3    	
	100 - 119  	4    	
	120+          	5 

Here, too, speed is already normalized to league average. You’ll note, however, that “average”—100—ranks pretty high on the scale. That’s because the pool of baseball’s most prolific speedsters is small and the distribution of stolen bases does not form a normal bell curve.

The third character measures expected batting average.

	xBA          	MM
	.000 - .239   	0
	.240 - .254   	1
	.255 - .269   	2
	.270 - .284   	3
	.285 - .299   	4
	.300+         	5

This is where we start running into trouble. Given the decline in batting averages—and xBA as well—assigning scores based on raw data does not work anymore. Check out the distribution of players in 2014 under the existing table:

	xBA          	MM	Pct.
	.000 - .239   	0	34.0%
	.240 - .254   	1	22.6
	.255 - .269   	2	20.9
	.270 - .284   	3	13.8
	.285 - .299   	4	6.3
	.300+         	5	2.4

More than 77% of batters earned a MM score of 0, 1 or 2. Only 22.5% earned scores of 3, 4 or 5. We need to find a better balance.

If we index xBA levels to league average, we can create a much more useful table:

	xBA Index 	MM	Pct.
	0-87   		0	16.3%
	88-92   	1	13.3
	93-97   	2	19.7
	98-102	   	3	17.0
	103-107   	4	17.7
	108+         	5	16.0

Now, 49.3% of batters earn a MM score of 0, 1 or 2 and 50.7% earn a score of 3, 4 or 5.

The fourth character measures playing time. This can remain unchanged.

        Role 			PA           	MM
	Potential full-timers	450+    	5
	Mid-timers		250-449    	3
	Fringe/bench		100-249   	1
	Non-factors		0-99		0


The first character in the pitching MM code measures xERA, which captures a pitcher's overall ability and is a proxy for ERA, and even WHIP. Once again, pitching dominance completely skews the distribution in the current chart.

	xERA         	MM	Pct.
	4.81+         	0	3.1%
	4.41 - 4.80   	1	9.8
	4.01 - 4.40   	2	19.3
	3.61 - 4.00   	3	23.9
	3.21 - 3.60   	4	23.4
	3.20-         	5	20.6

More than two thirds of pitchers earned a MM score of 3, 4 or 5. Only 32.1% earned scores of 0, 1 or 2. We need to find a better balance. Similar to speed, this skill is distributed in such a way that it’s impossible to create a normal bell curve. Even normalizing to league average can’t flatten the distribution. But we can come close.

	xERA Index   	MM	Pct.
	0-80         	0	5.9%
	81-90  	 	1	19.8
	91-100   	2	26.0
	101-110   	3	18.5
	111-120   	4	14.9
	121+         	5	14.9

While it’s nearly an even split between the top three scores and bottom three scores, you can see that there is still imbalance within the bottom group.

The second character measures strikeout ability. And again, a gross imbalance.

	K/9    		MM	Pct.
	0.0 - 5.3   	0	7.5%
	5.4 - 6.3   	1	15.2
	6.5 - 7.3   	2	20.6
	7.4 - 8.3   	3	21.3
	8.4 – 9.3   	4	13.1
	9.4+     	5	21.9

And fixing it as best as we can:

	K/9 Index	MM	Pct.
	0-76   		0	15.9%
	77-88   	1	16.2
	89-100   	2	19.5
	101-112   	3	18.3
	113-124   	4	12.3
	125+    	5	17.7

The third character measures saves potential.

	Description					Saves est.	MM
	No hope for saves; starting pitchers		0		0
	Speculative closer				1-9		1
	Closer in a pen with alternatives		10-24		2
	Frontline closer with firm bullpen role	        25+	        3

The fourth character measures playing time.

	Role 			IP        	MM
	Potential #1-2 starters	180+      	5
	Potential #3-4 starters	130-179   	3
	#5 starters/swingmen	70-129   	1
	Relievers		0-69		0

The Portfolio3 Plan

This normalization process has to carry over when creating profiles for the Portfolio3 Plan. Integrating the Mayberry values helps.

First, a review.

The foundation of Portfolio3 was the assertion that, when it comes to profitability, all players are not created equal. Every player has a different role on your team by virtue of his skill set, dollar value/draft round, position and risk profile. When it comes to a strategy for how to approach a specific player, one size does not fit all.

We need some players to return fair value more than others. We rely on some players for profit more than others. We can afford to weather more risk with some players than with others.

We need a way to integrate all these different types of players, roles and needs. The Portfolio3 Plan provides a three-tiered structure to the draft. P3 advises to diversify your roster with three different types of players. Depending upon the stage of the draft (and budget constraints in auction leagues), P3 uses a different set of rules for each tier from which you’ll draft. The three tiers are:

1. Core Players

2. Mid-Game Players
3. End-Game Players

In the original P3 structure, we used our sabermetrics to help filter the player pool. Now that we need to shift away from the raw gauges, it becomes more useful to integrate Mayberry into the process. Thus:

The Mayberry Portfolio3 Plan (MP3)

Mayberry scores can be used as a proxy for the original Portfolio3 filters, and they make more sense now. Most of the below will not be new, but I’ve made several tweaks to fine-tune the process.


General Roster Goals
Auction target: Budget a maximum of $160. Any player purchased for $20 or more should meet the Tier 1 skills criteria.
Snake draft target: 5-8 players, with an emphasis on those drafted in the earlier rounds
Reliability grades: No worse than “B” for each variable (Health, Experience and Consistency)
Playing time: No restrictions, however, pricier early round players should have more guaranteed playing time
Batter skills: Minimum MM scores of 3 in xBA plus either PX or RSpd
Pitcher skills: Minimum MM scores of 3 in xERA and K/9

Tier 1 players provide the foundation to your roster. These are your prime stat contributors and where you will likely invest the largest percentage of your budget or early round picks. There is no room for risk here, so the majority of these core players should be batters.


General Roster Goals
Auction target: Budget between $50 and $100; players should be under $20
Snake draft target: 7-13 players
Reliability grades: No worse than “B” for Health, no worse than “C” for Experience and Consistency
Playing time: Must have a MM score of 5 for batters (meaning full-time batters) and minimum 3 for pitchers (meaning at least mid-rotation starting pitchers)
Batter skills: Minimum MM scores of 3 in xBA or PX or RSpd
Pitcher skills: Minimum MM score of 3 in xERA or K/9

Tier 2 is all about accumulating playing time, particularly for batters. If a player is getting 500 AB, he is likely going to provide positive value in Runs and RBI just from opportunity alone. And given that his team is seeing fit to give him those AB, he is probably also contributing somewhere else. Tier 2 pitchers help stockpile strikeouts or build your ERA foundation.


General Roster Goals
Auction target: Budget up to $50; players should be under $10
Snake draft target: 5-10 players
Reliability grades: No restrictions, except no Health grades of “F”.
Playing time: No restrictions
Batter skills: Minimum MM scores of 3 in xBA plus either PX or RSpd (same as Tier 1)
Pitcher skills: Minimum MM score of 3 in xERA

Tier 3 players are your gambling chips, but every end-gamer must provide the promise of upside. For that reason, the focus must remain on skill and conditional opportunity. MP3 drafters should fill the majority of their pitching slots from this group.

By definition, end-gamers are typically high risk players, but risk is something you’ll want to embrace here. If a Tier 3 player does not pan out, he can be easily replaced. As such, the best options should possess the MM skill levels noted above, and at least one of the following:

  • playing time upside as a back-up to a risky front-liner
  • an injury history that has depressed his value (but not chronically injured players)
  • solid skills demonstrated at some point in the past
  • minor league potential even if he has been more recently a major league bust 

Here is an updated list of players in each tier. One of the major benefits of the MP3 process is that any player failing to find a home in one of the tiers can be safely ignored. Either his skills are not draft-worthy or his risk-profile too dangerous, regardless of skill.

So that means you are not going to find players like Troy Tulowitzki (4255 FDD) and Matt Kemp (4245 FCD) in that draftable Tier 1 pool. As $25-$30 buys, their risk profile is too high to be considered at that price; players at that level need to be safer. Similarly, you are not going to find speculative upside plays in Tier 3 like Dylan Bundy (1201 FFF) or Robert Erlin (2203 DDB) either. We need to see better skills before we jump in.

By shrinking the draftable player pool, it makes the roster planning and construction process easier.


Click here to subscribe

  For more information about the terms used in this article, see our Glossary Primer.