Debating Past Moves / Developing a Framework for Analysis

[Farm]

(Almost?) everyone here is a fan of the Jays and we all have our opinions and analysis on moves that we make. The problem is that we don't have a record that we can go back on to analyze and improve our decision-making/opinion-making process with.

 

Would anyone here be interested in making an effort to develop a framework to analyze moves with? We spend so much time talking about moves and thinking about them that I thought it would be interesting if we made a group effort to improve our thought process. I was thinking we could begin by looking at past moves the Jays have made in recent history. For each transaction we can have two posters write a piece on why it was good and why it was bad. I think it would be cool to see if there are some methods of analysis that are better or worse and go on from there.

 

Definitely open to suggestions/improvements but just wondering if any of you would be interested in something like that.

[GD]

Would love to

[Governator]

Interesting, I wonder if I could code a site that scrapes all transactions and then get accumulated fwar per player to compare. Would be even better with steamer projections to analyze data at the time of any trade and directly compare it to the aftermath.

[Nafro]

Interesting, I wonder if I could code a site that scrapes all transactions and then get accumulated fwar per player to compare. Would be even better with steamer projections to analyze data at the time of any trade and directly compare it to the aftermath.

 

I think its awesome that you even know how to do this potentially. If it works even better. It would be nice if we could put together a math model for analyzing moves. What about fwar player has provided since transaction and also club record since transaction. I understand that club record has very limited dependence on any one particular player. Just something that would add to the discussion.

 

With Prospects we could use a spreadsheet like AG put together last year. Utilizing numerous ranking sites to give an overall ranking of prospects.

[King]

Interesting, I wonder if I could code a site that scrapes all transactions and then get accumulated fwar per player to compare. Would be even better with steamer projections to analyze data at the time of any trade and directly compare it to the aftermath.

 

I know Juiced was looking for ways to improve or add to the BJMB. If you could code it onto this website that would be interesting.

[Governator]

I know Juiced was looking for ways to improve or add to the BJMB. If you could code it onto this website that would be interesting.

 

Well if I were to do it, it'd be its own site to automate it league wide, ongoing.

[Laika]

Interesting, I wonder if I could code a site that scrapes all transactions and then get accumulated fwar per player to compare. Would be even better with steamer projections to analyze data at the time of any trade and directly compare it to the aftermath.

 

The guys who run NEIFI always do stuff like this and then tweet the results. They do it all in-house though, with black box methodology (that is probably trash.)

 

https://twitter.com/neifico?lang=en

 

example

 

 

If you have the underlying skills I imagine it would be relative easy to just scrape BREF and Fangraphs for Steamer projections and contract status, and then do some sort of simple age adjustments for future years, turn every asset into a $ value and compare.

[Governator]

The guys who run NEIFI always do stuff like this and then tweet the results. They do it all in-house though, with black box methodology (that is probably trash.)

 

https://twitter.com/neifico?lang=en

 

example

 

 

If you have the underlying skills I imagine it would be relative easy to just scrape BREF and Fangraphs for Steamer projections and contract status, and then do some sort of simple age adjustments for future years, turn every asset into a $ value and compare.

 

Connecting everything together and making sure you're pulling the correct players are where things get complicated, especially when data is coming from different sources. You want to pull the right Jose Bautista for example.

 

Connecting data to contracts and dates would be a challenge as well.

[Jimcanuck]

This would have similarities to predicting pavement performance and investment returns.

[GD]

Connecting everything together and making sure you're pulling the correct players are where things get complicated, especially when data is coming from different sources. You want to pull the right Jose Bautista for example.

 

Connecting data to contracts and dates would be a challenge as well.

 

If you want help on this I probably can around when the season starts up

 

Have good familiarity with Python at this point

[SAAviour]

Would be cool even if it were very basic....total of $ each team took on, and total of WAR added up amongst traded players over life of contract, with maybe a couple more aspects.

Some trades would still be outstanding on winner/loser depending on how long each players contract is still for. ie Donaldson trade would not have a conclusion until the SS we gave up has done his 6 years of service with the As (mind block on his name right now)....but the Vernon Wells trade would have a conclusive winner vs loser.

 

Problem would be extensions...to include the Dickey extension or not for instance. I would think you would if the extension was immediate, but otherwise no.

[Governator]

Would be cool even if it were very basic....total of $ each team took on, and total of WAR added up amongst traded players over life of contract, with maybe a couple more aspects.

Some trades would still be outstanding on winner/loser depending on how long each players contract is still for. ie Donaldson trade would not have a conclusion until the SS we gave up has done his 6 years of service with the As (mind block on his name right now)....but the Vernon Wells trade would have a conclusive winner vs loser.

 

Problem would be extensions...to include the Dickey extension or not for instance. I would think you would if the extension was immediate, but otherwise no.

 

 

Using just WAR totals to compare trades obviously doesn't show the full picture but it would be a good starting point to get something going that we can add to it later.

 

A good example is the Adam Lind trade for Marco Estrada both with 1 full seasons going both ways.

 

On the surface, this seems like a very simple comparison, Adam Lind generated 2.3 fWAR for the Brewers while Estrada did 1.8 for the Jays. The trade would show a 0.5 win for the Brewers but Estrada is one of those special FIP beaters and put up a 3.13 ERA in 2015 that goes unrecognized in his WAR that's based on FIP. So I think more has to be considered when creating the calculation to be fair to all types of players where not everything gets shown in their WAR total. Jays also extended or at least re-signed him as a free agent afterward that likely wouldn't have happened if he wasn't already a Jay but I don't think that can be included here, it wouldn't be fair to Adam Lind or the Brewers.

[SAAviour]

Good point, and I had no idea Lind put up 2.3 fwar that year!

 

This point made me think of another idea...somehow add a poll to the site where fans can vote who the winning trade was....just for fun and to bring people into site.

[Governator]

Another snag in this is having to track flipped pieces, which would have to be a manual process I think to update trades. Think Roy Halladay trade, flipped Wallace for Gose, eventually flipping him for Travis. That's all part of the overall trade value that needs to be considered.

[Brownie19]

I'd just give up personally...

[Governator]

I'd just give up personally...

 

That's why you suck at fishing too

[Nafro]

Using just WAR totals to compare trades obviously doesn't show the full picture but it would be a good starting point to get something going that we can add to it later.

 

A good example is the Adam Lind trade for Marco Estrada both with 1 full seasons going both ways.

 

On the surface, this seems like a very simple comparison, Adam Lind generated 2.3 fWAR for the Brewers while Estrada did 1.8 for the Jays. The trade would show a 0.5 win for the Brewers but Estrada is one of those special FIP beaters and put up a 3.13 ERA in 2015 that goes unrecognized in his WAR that's based on FIP. So I think more has to be considered when creating the calculation to be fair to all types of players where not everything gets shown in their WAR total. Jays also extended or at least re-signed him as a free agent afterward that likely wouldn't have happened if he wasn't already a Jay but I don't think that can be included here, it wouldn't be fair to Adam Lind or the Brewers.

 

http://www.fangraphs.com/fantasy/finding-pitching-sleepers-with-infield-fly-rate/

 

This helps with FIP beaters, I think. I'm not sure if a IFB component to calculation could be added, to our calculation.

[Brownie19]

That's why you suck at fishing too

 

Touche. We'll played