New York, New York : A Tale of two Teams

On July 24th, the talk of the baseball world was how the New York Mets were a big success while the Yankees were closer to last place than first

Team	Division Race	Record	Playoff Chances
Mets	1st Place	51-43	63%
Yankees	3rd Place	50-44	35%

40 days later, the Mets odds of a post season appearance plunged to 1%. The Yankees chances jumped to 98%.

Team	Division Race	Record	Playoff Chances
Mets	3rd Place	61-67	1%
Yankees	2nd Place	76-52	98%

Lets examine what happened by looking at the daily change in ELO scores with each game.

In this excercise, we will stick with Base R.

Retrieve data from fivethirtyeight.com

mlb_elo <-read.csv("https://projects.fivethirtyeight.com/mlb-api/mlb_elo.csv")

2. Isolate the mets and yankees in the past month

mlb_elo_2 <- subset(mlb_elo,(team1=="NYM" | team2=="NYM" | team1=="NYY" | team2=="NYY" ) & date < "2021-08-28" & date > "2021-07-28" )[c("date","team1", "team2", "elo1_post" ,"elo2_post","score1", "score2" )]

3. Rearrange the data into 2 tables to make it more readable i.e. the Mets table would look like

Date	Opponent	Outcome	ELO Rating
2021-08-27	WSN	L	1490.045
2021-08-26	SFG	L	1491.586

mets_away <- subset(mlb_elo_2,team1=="NYM")[c("date", "team2", "elo1_post", "score1", "score2" )]
names(mets_away)<-c("date", "opponent", "elo", "score_Mets", "score_Opponent" )
mets_away$outcome  <- ifelse(mets_away$score_Mets > mets_away$score_Opponent, "W", "L")
mets_away$loc="A"




mets_home <- subset(mlb_elo_2,team2=="NYM")[c("date", "team1", "elo2_post", "score2", "score1" )]
names(mets_home)<-c("date", "opponent", "elo", "score_Mets", "score_Opponent" )
mets_home$outcome  <- ifelse(mets_home$score_Mets > mets_home$score_Opponent, "W", "L")
mets_home$loc="H"


mets<-rbind(mets_away, mets_home)

   # sort by date and apply a reverse index
mets<-mets[order(mets$date, decreasing =FALSE), ] 
rownames(mets)<-seq(28:1)


knitr::kable(mets, caption='Mets : What Happened in the past Month ??')

Mets : What Happened in the past Month ??
date	opponent	elo	score_Mets	score_Opponent	outcome	loc
2021-07-29	ATL	1509.421	3	6	L	A
2021-07-30	CIN	1506.525	2	6	L	A
2021-07-31	CIN	1507.809	5	4	W	A
2021-08-01	CIN	1504.194	1	7	L	A
2021-08-02	FLA	1501.963	3	6	L	H
2021-08-03	FLA	1500.574	4	5	L	H
2021-08-04	FLA	1502.547	5	3	W	H
2021-08-05	FLA	1500.728	2	4	L	H
2021-08-06	PHI	1499.111	2	4	L	H
2021-08-07	PHI	1497.515	3	5	L	H
2021-08-08	PHI	1495.595	0	3	L	H
2021-08-10	WSN	1496.869	8	7	W	A
2021-08-12	WSN	1500.131	5	4	W	A
2021-08-12	WSN	1498.891	4	1	W	A
2021-08-13	LAD	1499.040	5	6	L	A
2021-08-14	LAD	1497.959	1	2	L	A
2021-08-15	LAD	1494.506	4	14	L	A
2021-08-16	SFG	1493.229	5	7	L	H
2021-08-17	SFG	1492.264	2	3	L	H
2021-08-18	SFG	1495.823	6	2	W	H
2021-08-19	LAD	1494.496	1	4	L	H
2021-08-20	LAD	1493.655	2	3	L	H
2021-08-21	LAD	1492.822	3	4	L	H
2021-08-22	LAD	1497.312	7	2	W	H
2021-08-24	SFG	1493.922	0	8	L	A
2021-08-25	SFG	1492.748	2	3	L	A
2021-08-26	SFG	1491.586	2	3	L	A
2021-08-27	WSN	1490.045	1	2	L	A

Do the same for the Yankees

yankees_away <- subset(mlb_elo_2,team1=="NYY")[c("date", "team2", "elo1_post", "score1", "score2" )]
names(yankees_away)<-c("date", "opponent", "elo", "score_Yankees", "score_Opponent" )
yankees_away$outcome  <- ifelse(yankees_away$score_Yankees > yankees_away$score_Opponent, "W", "L")
yankees_away$loc="A"




yankees_home <- subset(mlb_elo_2,team2=="NYY")[c("date", "team1", "elo2_post", "score2", "score1" )]
names(yankees_home)<-c("date", "opponent", "elo", "score_Yankees", "score_Opponent" )
yankees_home$outcome  <- ifelse(yankees_home$score_Yankees > yankees_home$score_Opponent, "W", "L")
yankees_home$loc="H"


yankees<-rbind(yankees_away, yankees_home)
yankees<-yankees[order(yankees$date, decreasing = FALSE), ] 

rownames(yankees)<-seq(28:1)


knitr::kable(yankees, caption='Yankees : What Happened in the past Month ??')

Yankees : What Happened in the past Month ??
date	opponent	elo	score_Yankees	score_Opponent	outcome	loc
2021-07-29	TBD	1525.869	0	14	L	H
2021-07-30	FLA	1527.681	3	1	W	H
2021-07-31	FLA	1529.469	4	2	W	H
2021-08-01	FLA	1531.233	3	1	W	H
2021-08-02	BAL	1526.726	1	7	L	A
2021-08-03	BAL	1530.044	13	1	W	A
2021-08-04	BAL	1532.319	10	3	W	A
2021-08-05	SEA	1533.717	5	3	W	A
2021-08-06	SEA	1534.776	3	2	W	A
2021-08-07	SEA	1535.824	5	4	W	A
2021-08-08	SEA	1533.616	0	2	L	A
2021-08-09	KCR	1535.031	8	6	W	H
2021-08-10	KCR	1532.001	4	8	L	H
2021-08-11	KCR	1533.775	5	2	W	H
2021-08-12	CHW	1532.420	8	9	L	H
2021-08-14	CHW	1534.394	7	5	W	H
2021-08-15	CHW	1536.342	5	3	W	H
2021-08-16	ANA	1537.357	2	1	W	A
2021-08-17	BOS	1540.442	2	0	W	A
2021-08-17	BOS	1538.910	5	3	W	A
2021-08-18	BOS	1542.302	5	2	W	A
2021-08-19	MIN	1543.577	7	5	W	A
2021-08-20	MIN	1546.327	10	2	W	A
2021-08-21	MIN	1548.568	7	1	W	A
2021-08-23	ATL	1551.461	5	1	W	H
2021-08-24	ATL	1552.961	5	4	W	H
2021-08-26	OAK	1554.381	7	6	W	H
2021-08-27	OAK	1557.711	8	2	W	H

Now plot the progression of the elo scores.

# set the y limit to the lowest and highest ELO score

plot(yankees$elo, type="l", ylab="ELO", xlab="Day", col="blue", ylim=c(1488,1560), main="ELO : One Month in New York Baseball")   
lines(mets$elo, col="red", lty=1)
legend("topleft", legend=c("Mets", "Yankees"), col=c("red", "blue"), lty=1, cex=0.8,text.font=4, bg='lightblue')

Data607_Week1

Tom Buonora

August 28, 2021

New York, New York : A Tale of two Teams

On July 24th, the talk of the baseball world was how the New York Mets were a big success while the Yankees were closer to last place than first

40 days later, the Mets odds of a post season appearance plunged to 1%. The Yankees chances jumped to 98%.

Conclusion

Both teams now have a record that conforms to preseason expectations so we can use the law of large numbers to explain the change

The Mets recent record can be explained in part to the loss of their best pitcher plus a brutal road trip against some very good teams

The Yankees on the other hand, as evidenced by a profound improvement in their ELO score simply got hot and are currently playing at a high level