Part of the title of this FanPost ("It's Gut Check Time!") is inspired by one of Amazin' Avenue's orginal trolls, centerpede 101, whose contributions have earned him a spot in the Amazin' Avenue Troll Hall of Fame. Do read some of his "material"- it's pure, um... pyrite. So, for you aspiring trolls out there, here's a father figure to lean on. I'll bet that somewhere, centerpede 101 is smiling/scowling down upon us from troll heaven.
To quote a phrase from him, June is "Gut Check Time". The tradition continues.
Often times here at Amazin' Avenue, just like many of you, I constantly endure the ebbs and flows of the team, and with it, the schizophrenic nature of the valuation of our team from very temperamental people. It can only take a few games for people to say "This team is going places!" to "We suck, we are totally hopeless, **** this team!" Unlike many of these types, however, I for one understand that the strength of the team is constantly changing, and that a few games do not dictate whether or not a team is good or bad. Honestly, some of the crap I see on the game threads just makes me shake my head.
Introducing a new way to evaluate our team: MOMENTUM.
Theoretically speaking, the team will always be at least a little better/worse after a game. A LITTLE. The strength of the team, just like any other team, varies over time. Usually, when we speak of how good a team is, we speak of the present, and not the team throughout the year. It's often hard to differentiate a team or player's progress when we look at the year-to-date stats, because taking the aggregate stats do not tell the whole story.
Therefore, I have developed an extremely complicated formula for measuring the momentum of the team, which is based on two things: recency and number of instances. In this post, I will mull over the team's offense throughout the season. wRC+ (normalized weighted runs created), which is park and league adjusted, is the best way to evaluate offensive production. It removes statistical noise caused by luck and variances in ballpark eccentricities, to measure the true strength of an offense.
Here is a graph and table of the team's game-by-game wRC+ throughout the 60 games that have been played so far (click to embiggen):
As you can see, the team is very inconsistent on offense, and cannot establish a discernible, coherent pattern. However, I have found a way to adjust the accumulated wRC+, so that it can be weighted based on recency and PAs (excluding intentional walks). All the data that is needed to create this is in the above table. To set it up, so that the weights of the data by date decay to 50% of the weight of the current date in 14 days:
a) First, for every day elapsed from a chosen date in which a game has occurred, a weight factor of 0.949951607083599 is multiplied to produce a weight.
b) Then all of these weights are added, so that a sum is produced. (Sum X)
c) To factor in adjusted PAs,
- each of the adjusted PAs are multiplied by the wRC+ for each of the corresponding dates, and then multiplied again by the weights of the corresponding dates, based on a chosen date. (The corresponding date must be at or prior to the chosen date.) The sum of these values up to a chosen date, we will call sum A.
- Then, each of the adjusted PAs are multiplied by 100 (which is representative of the average wRC+), and then multiplied again by the weights of the corresponding dates, based on a chosen date. The sum of these values up to a chosen date, we will call sum B.
- Based on a chosen date, to get the momentum-adjusted wRC+, just divide sum A by sum B, and then multiply by 100. This is the result, based on the chosen date.
However, it is important to point out that a buffer must be placed into the data, so that for chosen dates close to or at the start of the season, the wRC+ results do not cause wild swings in the momentum shown. For this, 14 games worth of "filler statistics" are created prior to game 1, with the current league-average amount of adjusted PAs for each of those days (in this case, 37.9652855543112), and a wRC+ input of 100 for each of those days. Then, we just treat these filler stats as if they were part of the sample of the data.
Based on this process, this is what we get:
Because the momentum-adjusted wRC+ is based for each chosen date on the variations created by each game played, as well as a decay in the impact of older games, a continuous line can be placed to connect all of the points together. Think of this as a "stock market" trend or valuation of the Mets offense. With that said, a momentum consistently deviating towards well below 100 denotes a slump (see 4/16 - 4/23 for a perfect example), and a momentum doing the opposite denotes a hot streak (see 5/9 - 5-13).
In order to better analyze all of this data, here's an additional chart and table to enjoy:
Stick around for the It's Gut Check Time! series, where I will apply the methodology I just presented here to evaluate our team, and where we're headed (all stats will be as of 6/5/2014).