Welcome to the longest running online
fantasy baseball game and community!

Obie at 4:08p Jan 21

The original post has been updated. Tom Tango analyzed my data and noted that my original correlation comparisons had mixed rate stat and counting stat results. This problem has now been corrected. He also noted that the BBM offensive formula is actually the original Bill James Runs Created formula. This revelation will be covered in more detail within a future post.

Obie at 11:39p Jan 08

That is exactly why I like the Baseball Manager Progression leagues with Pinetar. Everyone fights to the end because they want to either get promoted or keep from being demoted. As you advance with promotion, you are also challenged with stiffer competition from managers that have proven themselves at lower tiers. If you really want to be the best, you have to beat the best. If you just want stiff competition every year, that is exactly what your going to get.

EWJ at 10:39a Jan 06

Yeah, it looks like FIPs has the advantage when there weren't enough innings pitched for the luck in ERA to even out. But once there are enough innings pitched, which happens much sooner for starters, then ERA is slightly better. And the greater the number of innings pitched, such as starters in MORPS, the greater the advantage of using ERA over FIPS.

Obie at 1:01p Jan 06

I have to think that ZiPS already uses some type of reliability calculation within their projections. That is how you typically regress to the mean for most projection systems. I've been working with historical data over the last several days to see if I could somehow "crack the formula" that ZiPS uses for their batters. I believe the fact that he models the data for each team and doesn't distinguish between Major League and Minor League ABs accounts for better projections on plate appearances. It doesn't really help fantasy owners because we only care about MLB ABs, but it does make it easier to project. The one that I don't get is the ZiPS accuracy on stats per plate appearance. MORPS beat ZiPS with pitchers. MORPS didn't really come close with batters. One thing I determined late yesterday was that breaking down the stats by first half and second half allows you to give more strength to second half stats going into the coming year. When I applied this to historical data, MORPS made up some significant ground on ZiPS. I believe that their are some correlations that I have not found yet that may make up the rest.

EWJ at 10:05a Jan 06

So perhaps using ZiPS with an adjustment for reliability would be a good way to go?

Obie at 12:46p Jan 06

Interesting!!! If I get some time over the next couple of days, I'll try to run some regression tests on a couple of these stats compared to actual 2011 performance. I believe we'll find that their are strong correlations between historical K and BB rates and 2011 performance. That is probably why FIP was created. It includes HRs, BBs, Ks and innings pitched.

Obie at 12:21p Jan 06

The anomaly with ERs, Runs and HRs was really only seen with relievers. Starters did have the lowest number of HRs allowed over the last 5 years, but it was not a dramatic shift. MORPS projections were off by 2.4%, 1.6% and 2.2% in each of these areas respectively for starters. Relievers were another story. MORPS projections were off by 12.3%, 11.5% and 8.1% in each of these areas respectively for releivers. For the top 175 relief pitchers, ERs were down 13.7% from the 4 year average, Runs were down 12.7% from the 4 year average and HRs were down 9.2% from the 4 year average. It just seems like 2011 was a huge year for top relievers. ERA and FIP were at 5 year lows. BABIP and WHIP were at 5 year lows. Ks per 9 innings were at a 5 year high and BBs per 9 innings were at 5 year lows. I just can't see that happening again. I know we are in an era of specialization, but there is no way we can expect relievers on whole to do as well as they did last year.

EWJ at 9:29a Jan 06

I really don't have a good projection system for pitchers. (Well, I don't really have a projection system at all. What I mean is what data I look at when I'm making my decision what to project from a guy isn't that great.) For pitchers, the last two years I looked at a combination of two things. First of all I took pitcher's K/BB rate and multiplied it by IP to get a pitcher's rough theoretical value and put them in order from best to worst. (On the questionable presumption that K/BB is a better indicator of how good a pitcher they are than ERA. It is an extremely questionable as a way to rate guys considering it in no way takes into account whether the guy actually gives up a ton of hits or very few) The second thing I look at is the pitcher's batting average on balls in play, and compare it to the previous 2 years BABIP. Then I look for guys with good theoretical value that had a higher BABIP than they normally do last year. My presumption there is that these guys were just unlucky, and will be undervalued in the draft, so I targeted them by moving them up on the list. Guys the were lucky last year (lower BABIP than normal) I moved down. Yes, overly, stupidly, simple. But the last 2 years using that system I had ERAs in my Pegasus league of 2.98 and 3.10, while the previous two years using my old system I had ERAs of 3.88 and 3.69. So, while it might not be a good system, so far it looks like it's at least a better system than my old one. (I used to rely heavily on ERC, component ERA, looking for value, and it just didn't work) While my team led Armageddon in ERA, I think that had more to do with my strategy of just chasing the hot stat pool from the free agent list than it had to do with my drafting. I don't think many of my drafted pitchers stuck with the team over the long term in that league.

EWJ at 9:09a Jan 06

If the trend for better relief pitching continues, then eventually MORPS will catch up to it once you have 4 years of data of the trend. And between now and then it will get progressively closer as more of the years used in the calculation occurred under this trend. When you say last year had the lowest number of runs, earned runs, and homers allowed, was that just by relievers or was that for all pitchers? If it was for all pitchers, then the question isn't why your relief projections were off, they were off because it was a great year for pitchers. The question is why your starting pitcher projections were correct when it was a great year for pitchers. Shouldn't starting pitcher projection have been off too?

Obie at 12:24p Jan 06

Yes - a rookie that has no MLB experience would project at the MORPS "adjusted" mean for their position. I found that using all players to compute the positional means actually adjusts projections for better players more than it should. There are plenty of players each year that get 50 or less ABs that will never make a MLB roster the rest of their lives. The MORPS positional means are adjusted to include only those players with more ABs than the minimum. Although I haven't discussed it on the blog yet, I actually calculate the means against left handed and right handed pitching and use that combination to project positional means overall. Getting back to the rookies, the mean is calculated at each stat per plate appearance. Based upon the modeling I will do this year of each team during spring training, some rookies will get more plate appearances based upon projected playing time. Thus their stats will adjust up or down based upon those projections. The thing that will not change is the per plate appearance positional mean for each specific stat. If a rookie is good enough to take a NL 1B job out of spring training, the assumption will be that they will be good enought to hit at the NL 1B positional mean rate multipled by their projected plate appearances.

EWJ at 9:23a Jan 06

"By calculating the reliability of a player’s projection using only major league data, MORPS adds a proportional dosage of a player’s positional mean to complete a rookie’s player projection." So does this mean a rookie with no major league experience would automatically be projected to perform at the mean for that position? I suppose that would be as good of a starting point as any. It would require some human thought to decide whether the projection the formula arrived at for that player seemed realistic, but that's always true for any mathematical projection system.

Obie at 9:14p Dec 19

You obviously didn't need my rankings last year - what a race in Armageddon. Hopefully #1 and #2 can both promote this year. I'm just glad I was finally able to win one. I've lost count of how many times I've taken 2nd place.

EWJ at 10:16a Dec 19

Oh, sure. Why couldn't you have done this last year so I knew what you were up to? ;-)