Beat the Streak Picks

Here are the top picks my program produced for use in Beat the Streak. This post mostly explains the ideas behind the calculations. In addition, this post shows tests on the Neural Network (NN).

First, the log5 list:

0.346 — Daniel Murphy batting against Jeremy Hellickson
0.325 — Eduardo Nunez batting against Kendall Graveman
0.324 — Jean Segura batting against Collin McHugh
0.317 — Marcell Ozuna batting against Jeff Locke
0.312 — Martin Prado batting against Jeff Locke
0.307 — Buster Posey batting against Mike Foltynewicz
0.307 — Starling Marte batting against Justin Nicolino
0.306 — Josh Harrison batting against Justin Nicolino
0.305 — Melky Cabrera batting against Matt Harvey
0.304 — Hunter Pence batting against Mike Foltynewicz

Now, the NN list:

0.307, 0.757 — Buster Posey batting against Mike Foltynewicz.
0.286, 0.749 — Jose Altuve batting against Edwin Escobar.
0.346, 0.746 — Daniel Murphy batting against Jeremy Hellickson.
0.300, 0.744 — Francisco Lindor batting against Colby Lewis.
0.306, 0.742 — Josh Harrison batting against Justin Nicolino.
0.325, 0.741 — Eduardo Nunez batting against Kendall Graveman.
0.286, 0.739 — Joe Panik batting against Mike Foltynewicz.
0.255, 0.739 — Ben Revere batting against Jeremy Hellickson.
0.283, 0.737 — Matt Duffy batting against Mike Foltynewicz.
0.287, 0.736 — Denard Span batting against Mike Foltynewicz.

I removed Dee Gordon and Michael Brantley from the NN list as they are not available to play.

If you look at the three-year stats for today’s starting pitchers, you see that Mike Foltynewicz has the double whammy of a very high BABIP with a very high home run rate. He is somewhat improved in 2016, but still is high in both categories.

The batter on the list that surprises me a bit is Ben Revere, who is off to a very slow start for the Nationals. With just 95 PA this season, his low Hit Average gets pulled a bit toward the MLB average. His previous two seasons were very good, so the NN is recognizing that his long term number is more important than his current year number at this point.

Once again, these lists are a guide. Use your own knowledge of ballparks, injuries, platoon advantages, etc. to make your picks.

from baseballmusings.com http://ift.tt/25wfTSX

Games of the Day

The Red Sox lead the Orioles by one game as Boston travels to Baltimore for a four-game series. Steven Wright takes the hill against Tyler Wilson. Wright is impressive for a knuckleball pitcher, as he doesn’t walk many batters. This year, he appears to be getting better movement, too, as opponents hit just three home runs off him, compared to 12 home runs in 12 more innings last season. Wilson is that rate pitcher who allows a high number of balls in play, but a low number of hits per inning. In his brief career, his ERA is well before his xFIP.

The Dodgers visit Chicago to play the best team in baseball, the Cubs. Alex Wood faces Jason Hammel. With good K/BB/HR numbers, Wood should have a better ERA than his 4.03 mark. He performed poorly with runners in scoring position, allowing a .359/.400/.513 slash line. Hammel’s 2.17 ERA would represent his best single season performance. He cut his home run rate more than 50% compared to the last two seasons.

Philadelphia find themselves just 3 1/2 games back in the NL East race as the first place Nationals come to town. Tanner Roark battles Jeremy Hellickson. Roark upped his strike out rate this season, and his walk rate rose with it. Trading hits for walks, however, is working well. Hellickson is eating innings, but is on a pace for about 30 home runs allowed. Most of those so far have come on the road.

Enjoy!

from baseballmusings.com http://ift.tt/24gJE4t

The White Sox Slide

The White Sox slipped to third place in the AL Central after a ten game span in which they lost five of six to Kansas City and three of four to Cleveland. For some reason, that put Robin Ventura‘s head on the block. What happened?

At the end of play on May 9th, the White Sox stood 23-10 with a six game lead in the AL Central. It looked like it would be the summer of Chicago baseball as both Windy City teams opened up nice division leads. Since then however, the White Sox are just 4-13.

The offense was playing well through May 9th, with a .249/.324/.394 slash line. They weren’t hitting for average, but they were getting on base with some power. They are getting on base less during the slide, .249/.307/.383.

The problem is really the pitching. During the hot streak, they allowed opponents a .225/.293/.339 slash line. That would have been enough to compensate the offensive slide. Instead, the White Sox itchers getting hammered to the tune of .279/.337/.444. Note that the starts in that time have good strikeout and walk numbers, but home runs are hurting them. They also give up more hits that I would expect. I’m wondering if other teams just decided to attack the White Sox staff early, since they are around the plate so much? If they did, it’s working, since they are driving the starters from the game early, putting more pressure on the White Sox bullpen.

If the latter is true, that opponents made an adjustment to the White Sox staff, then maybe Ventura deserves some blame. One thing the manager and his staff should do is recognize an opponent’s adjustment, and counter it. We’ll see what changes now that Chicago was knocked off their perch.

from baseballmusings.com http://ift.tt/1Z8k1Bs

A New Noob and Summer Noob Schedule

Hi folks!

As some of you might have noticed yesterday, we have a new Noob on board - Necropocalypse - you can find out more about him on his page here.

With Necropocalypse on board, we now number 5 current Noobs so I thought I'd give a quick run down of our schedule as we head into summer.

Monday - Noobish News/Q&As

Tuesday - Shamikebab

Wednesday - A Closer Look

Thursday - Merkeysa

Friday - ArkMechanicus

Saturday - Brambleten

Sunday - Necropocalypse

We plan on everything bar the News and Q&As to be weekly, with News as required and Q&As on a regular, ongoing basis. This may not always be possible, but with 5 of us, I'm really happy to say we should now have almost daily fresh new content. This is something I'm quite proud that we're going to be able to achieve and will hopefully continue for a long time. 

Cheers, 

Bram

from Noobs and their paintbrush http://ift.tt/25tiOZe

Weekly Look at Offense

Week eight of the 2016 season proved to be the highest scoring week of the year, coming in at 9.35 runs per game. That was nearly a run per game higher than last week, at 8.37 runs per game. The 2016 season continues to run ahead of the 2015 season at the same point, 8.67 runs per game this year versus 8.32 runs per game through eight weeks in 2015. (All comparisons are through eight weeks in each season.) As has been true all year, home runs and walks are way up, there is nearly one more strikeout per game, and other hits (singles, doubles, and triples) are down slightly. Overall, there are more batters reaching base, and more long hits to drive them home.

I suspect this difference between the seasons my get bigger over the next few weeks. As you can see on the graph associated with this spread sheet, 2015 entered a ten week down period in week eight. If 2016 can avoid that, we are looking at much higher scoring this season.

from baseballmusings.com http://ift.tt/25vIqs1

Memorial Day Update

The Day by Day Database is up to date. My thoughts go out to the family and friends of those braves souls who gave their lives to protect their country.

from baseballmusings.com http://ift.tt/1smkTb7

More on Beating the Streak

I ran a test of the Neural Network (NN) I’m using to produce a set of output for the Beat the Streak game. I built the network by selecting 10 random dates from each even year from 1980 to 2008, a total of 15 years. I took all games on those 150 dates, determined the starting pitcher for each game, and the starting opposition lineup for each game, leaving out the opposing pitcher. My guess is that worked out to be about 13 games a day, 17 position players per game, or a total of about 33,000 samples. I then trained the network on 75% of those samples, and used 25% for validation. I ran 200 training epochs, but the NN converged rather quickly.

To test, I used 160 dates, 40 each from 2003, 2005, 2007, and 2009. So there is no overlap with the training data. For each date, I used the NN to determine the match-up with the highest probability of getting a hit, and if the batter did indeed get a hit that day. I ordered the data by date, and looked for streaks. Here are the results:

Days: 160, Expected Games with hit: 124.9, Actual Game with hit 129.
Streak Length: 1, Number of times: 8
Streak Length: 2, Number of times: 5
Streak Length: 3, Number of times: 2
Streak Length: 4, Number of times: 3
Streak Length: 5, Number of times: 2
Streak Length: 6, Number of times: 2
Streak Length: 8, Number of times: 1
Streak Length: 9, Number of times: 1
Streak Length: 10, Number of times: 1
Streak Length: 13, Number of times: 1
Streak Length: 14, Number of times: 1
Streak Length: 16, Number of times: 1

So in this case, the NN underestimates the probability of the batter getting a hit that day. That’s good, I prefer a conservative model. It predicts a probability of 0.78, and delivers .806. It’s in the ball park.

Note however, that the long streaks are not very long. The season only made it into double digits four times.

As a sanity check, here is what happens when the worst player is chosen every day:

Days: 160, Expected Games with hit: 94.4, Actual Game with hit 97.
Streak Length: 1, Number of times: 16
Streak Length: 2, Number of times: 9
Streak Length: 3, Number of times: 9
Streak Length: 4, Number of times: 2
Streak Length: 5, Number of times: 1
Streak Length: 7, Number of times: 2
Streak Length: 8, Number of times: 1

The NN expects with a probability of .59, and the players produce at a rate of .606 per game. Not only that, but there are no double digit streaks. I’m liking this a lot.

from baseballmusings.com http://ift.tt/1UdoA96