ESPN’s Bradford Doolittle thinks the O’s currently are a 72-win team

[btdart20]

4 minutes ago, sportsfan8703 said:

I don’t think people realize how much we got out of Voth and Watkins last season. Watkins threw 120 IP with a 4.3 ERA. Voth had a 10 ERA when we signed him and then pitched like one of the best SP in the AL in the 2nd half. So we did get a little lucky with those two SP. Grayson should help mitigate the regression there. I think adding Eovaldi or Wacha makes us a solid 85+ win team. 

Regression from Voth and Watkins types exactly what these prognosticators are seeing.  They aren’t buying into the changes being real.  They are likely right in direction for individual players.  But in degree?  In key players?  In frequency?  I think they are making some assumptions out of context.

[Frobby]

Here’s Doolittle’s description of his methodology:

“These numbers here are a rough cut. Underpinning them are the player projections from Steamer, available at Fangraphs.com. I've made tweaks to playing time forecasts on my depth charts in order to compile a basic win projection for each team, which I used as the basis for running 10,000 simulations of the 2023 schedule.”

The interesting thing about that is that I believe Fangraphs came up with 78-79 wins using the same Steamer projections.  

[Pickles]

1 minute ago, Frobby said:

Here’s Doolittle’s description of his methodology:

“These numbers here are a rough cut. Underpinning them are the player projections from Steamer, available at Fangraphs.com. I've made tweaks to playing time forecasts on my depth charts in order to compile a basic win projection for each team, which I used as the basis for running 10,000 simulations of the 2023 schedule.”

The interesting thing about that is that I believe Fangraphs came up with 78-79 wins using the same Steamer projections.  

And that was without Gibson, Frazier, Givens, and McCann.

[Frobby]

3 minutes ago, Pickles said:

And that was without Gibson, Frazier, Givens, and McCann.

Was it?   I know Szymborski did one using ZiPS before those transactions.   I thought the Steamer-based one was more recent and included those players, but I could be wrong.   In any event, if you are going to add in the players we acquired, you have to do the same for the other teams.   

[Pickles]

3 minutes ago, Frobby said:

Was it?   I know Szymborski did one using ZiPS before those transactions.   I thought the Steamer-based one was more recent and included those players, but I could be wrong.   In any event, if you are going to add in the players we acquired, you have to do the same for the other teams.   

I had thought they put those out in the beginning of November, almost immediately after the season, basically simultaneously.

Either way, I as you, am interested in how he took a projection system that projects the Os at 78-9 wins, and used it to project the Os at 72 wins.

Perhaps Steamer sees a lot more downside than upside in the Os and this reflects that.  Or perhaps it isn't as rigorous as it is claimed to be.

[Frobby]

5 minutes ago, Pickles said:

I had thought they put those out in the beginning of November, almost immediately after the season, basically simultaneously.

Either way, I as you, am interested in how he took a projection system that projects the Os at 78-9 wins, and used it to project the Os at 72 wins.

Perhaps Steamer sees a lot more downside than upside in the Os and this reflects that.  Or perhaps it isn't as rigorous as it is claimed to be.

At the end of the day, all these projections systems are just educated guesswork with huge margins of error.   I look at them for entertainment purposes but would never bet money based on what they say.   

Right now the Frobbometer projections have us winning 81 games, plus or minus 10.    
 

[Sports Guy]

29 minutes ago, Pickles said:

And that was without Gibson, Frazier, Givens, and McCann.

All moves sure to make teams quiver with fear!

[tntoriole]

32 minutes ago, Frobby said:

Here’s Doolittle’s description of his methodology:

“These numbers here are a rough cut. Underpinning them are the player projections from Steamer, available at Fangraphs.com. I've made tweaks to playing time forecasts on my depth charts in order to compile a basic win projection for each team, which I used as the basis for running 10,000 simulations of the 2023 schedule.”

The interesting thing about that is that I believe Fangraphs came up with 78-79 wins using the same Steamer projections.  

What is amazing is how wrong these predictions routinely are every year.  No better than Sporting News used to do 50 years ago.. with zero computer algorithms lol

[vab]

Tell me what their projection for 2022 was this time last year. 

[Number5]

I feel relatively certain that the Orioles will win between 40 and 120 games.

[Pickles]

11 minutes ago, Sports Guy said:

All moves sure to make teams quiver with fear!

While nobody is quivering, those moves undoubtedly make the team better.  

[Sports Guy]

Just now, Pickles said:

While nobody is quivering, those moves undoubtedly make the team better.  

That’s undoubtedly false.

[Pickles]

Just now, Sports Guy said:

That’s undoubtedly false.

On paper they're undoubtedly better with those additions.

I concede the games aren't played on paper.

[Pickles]

25 minutes ago, Frobby said:

At the end of the day, all these projections systems are just educated guesswork with huge margins of error.   I look at them for entertainment purposes but would never bet money based on what they say.   

Right now the Frobbometer projections have us winning 81 games, plus or minus 10.    
 

As someone else just stated, an educated fan using "common sense" and intuition is just about as good as any formulaic projection system.

Your projection is similar to my own, and I'd say there's about a 90% chance of the Os winning the amount of games in the provided range.

[maybenxtyr]

1 hour ago, tntoriole said:

Would you be surprised by 92? Lol 

Certainly! It would be awesome, but definitely unexpected.