I ran a multi-variable linear regression analysis
Nobody will careI ran a multi-variable linear regression analysis on the 2049 data vs OSA ratings for all batters who played at least 40 games. I analyzed the following variables and how they relate to WAR:
Contact
Gap
Power
Eye
Ks
Defense (calculated based on listed position, I believe)
Speed
Steals
Base running
So before I reveal the information in a huge press release, I ask Utopian WBL owners the following questions:
1. Rank these ratings in order of importance.
2. Are any of them so irrelevant that they wash out completely when compared to more important variables?
I ran a multi-variable linear regression analysis on the 2049 data vs OSA ratings for all batters who played at least 40 games. I analyzed the following variables and how they relate to WAR:
Contact
Gap
Power
Eye
Ks
Defense (calculated based on listed position, I believe)
Speed
Steals
Base running
So before I reveal the information in a huge press release, I ask Utopian WBL owners the following questions:
1. Rank these ratings in order of importance.
2. Are any of them so irrelevant that they wash out completely when compared to more important variables?
This is awesomeI ran a multi-variable linear regression analysis on the 2049 data vs OSA ratings for all batters who played at least 40 games. I analyzed the following variables and how they relate to WAR:
Contact
Gap
Power
Eye
Ks
Defense (calculated based on listed position, I believe)
Speed
Steals
Base running
So before I reveal the information in a huge press release, I ask Utopian WBL owners the following questions:
1. Rank these ratings in order of importance.
2. Are any of them so irrelevant that they wash out completely when compared to more important variables?
I ran a multi-variable linear regression analysis on the 2049 data vs OSA ratings for all batters who played at least 40 games. I analyzed the following variables and how they relate to WAR:
Contact
Gap
Power
Eye
Ks
Defense (calculated based on listed position, I believe)
Speed
Steals
Base running
So before I reveal the information in a huge press release, I ask Utopian WBL owners the following questions:
1. Rank these ratings in order of importance.
2. Are any of them so irrelevant that they wash out completely when compared to more important variables?
Gooksta got me on the avoid KsI am OU but I value eye more than he does.
Gooksta got me on the avoid Ks
Avoid k with high contact and good gap is lethal imoWith Richmond I had a guy that was like 45/50 contact, poor eye, and then 80 avoid Ks. He wasn't replacement level I don't think.
I'm not trying to prove a point with that, it just popped in my head.
Speed plays a role in both thoughsteals dont matter at all when compared to baserunning. Steal 100 bases but get caught 100 times and you still suck.
Speed doesnt matter at all when compared to defense and baserunning either.
I personally rank
Contact
Power
Defense
Ks
Eye
Baserunning
Gap
Gooksta posits that with the above average stuff most starters have and the favor they garner with GMs over low stuff high movement, that Ks are more important now than ever. I agree with him on that. Walks are very important, it's hard to rank them. It's why i just trade for guys with all 5 toolsWith Richmond I had a guy that was like 45/50 contact, poor eye, and then 80 avoid Ks. He wasn't replacement level I don't think.
I'm not trying to prove a point with that, it just popped in my head.
But it isnt worth ranking if it doesnt translate. 80 speed but cant steal at a good clip and sucks at defense? Cool. That's why using the end stats and not the influential stats is my preferenceSpeed plays a role in both though
Less strikeouts = more batted balls in play. Which can lead to more productive outs/and more hits. Downside, could lead to more double playsSo what's the theory on avoid Ks?
They are the worst without a decent BABIPSo what's the theory on avoid Ks?
The problem is without speed many of those guys don't play to their position or defensive ratings.But it isnt worth ranking if it doesnt translate. 80 speed but cant steal at a good clip and sucks at defense? Cool. That's why using the end stats and not the influential stats is my preference
Contact and power are obviously way more important.It seems like everyone has a pretty decent fundamental idea of what is important, which really shows what great computers our brains are that they are essentially doing the same regression stuff on the fly. Karl got the top 3 right, OU was pretty close across the board. and Mr Radpants made a good observation about how he values eye more.
For the 2049 data set, here are your most important variables in the correlation between ratings and WAR:
1. Contact
2. Power
3. Eye
4/5. Ks/Defense (basically tied statistically)
6. Gap
Base running, Speed, and Stealing all washed out compared to the others. That isn't to say they aren't important for individual players, just that there are a lot of successful players who don't have good attributes in these areas and a lot of unsuccessful players who do. I think with speed/stealing being related to youth at times could factor in to why this washes out.
Anyone care to take a guess how important contact and power are compared to everything else?
Productive outs can lead to runs over a strikeout. A pitcher who can have a 10-12 strike out game that is close to 50% of unproductive outs in a game.. that just kills offensive production.They are the worst without a decent BABIP
But if they have good defensive stats it doesnt matter what their speed is. Also speed has no bearing on steal rating either, not even a little bit. Might on production of course.The problem is without speed many of those guys don't play to their position or defensive ratings.
That's what I'm saying. They don't have good defensive stats without the speed, even if their defensive ratings are good. I'm talking about having bad speed too, not saying they need 80.But if they have good defensive stats it doesnt matter what their speed is. Also speed has no bearing on steal rating either, not even a little bit. Might on production of course.
I didnt know travis was doing ratings, i thought it was based on stats. If the stats show he can run and play defense then you dont care how fast he is
Most are fast that can do both, but some arent
NoProductive outs can lead to runs over a strikeout. A pitcher who can have a 10-12 strike out game that is close to 50% of unproductive outs in a game.. that just kills offensive production.
I'll take a ground out that advances a runner over a strikeout any day of the week.
But it isnt worth ranking if it doesnt translate. 80 speed but cant steal at a good clip and sucks at defense? Cool. That's why using the end stats and not the influential stats is my preference
Right, avoid k with high contact and great speed and overall baserunning ability can be a game changer.. why I am so excited for rocky and TurtleContact and power are obviously way more important.
I also think baserunning would be more important than you think. For ages in the ubl/wbl you couldnt get guys to have positive BsRs because of the way steals were coded. Obviously that's fixed now and i had great success with it. As more people start to steal i think you'll see the value go up.
Mosterd is good iirc. Ive also had low speed guys have high baserunning and steal well enough. That's what im saying, speed doesnt always translate making it a bad marker. What does translate is looking at their defense, their xbh, and their stealsThat's what I'm saying. They don't have good defensive stats without the speed, even if their defensive ratings are good. I'm talking about having bad speed too, not saying they need 80.
Gooksta is right in looking for certain combinations that play well too. That's one thing that you miss when you just throw the whole mess together in one pot and perform a regression analysis. Speed/stealing/base running probably matter a lot more for your leadoff types (looking for high OBP) than they do for your power types.
Gooksta posits that with the above average stuff most starters have and the favor they garner with GMs over low stuff high movement, that Ks are more important now than ever. I agree with him on that. Walks are very important, it's hard to rank them. It's why i just trade for guys with all 5 tools
And it's my favorite that's why we made for good trade partners.If I didn't have what I have, I would build on types. It's what I look for in guys who don't have a tool or two. Gooksta has probably my favorite type in 3-4 of his prospects with his contact/k/steal combo.
The combo I hate the most is average contact high power.
I'm talking more about defense, as I don't pay attention to steals close enough. I think if you have 25 speed your defensive stats are shit regardless of ratings. I'm sure there are exceptions, but I have seen that trend.Mosterd is good iirc. Ive also had low speed guys have high baserunning and steal well enough. That's what im saying, speed doesnt always translate making it a bad marker. What does translate is looking at their defense, their xbh, and their steals
Btw baserunning doesnt have anything to do with steals either. It's only for the initial hit and once your on base
I'm talking more about defense, as I don't pay attention to steals close enough. I think if you have 25 speed your defensive stats are shit regardless of ratings. I'm sure there are exceptions, but I have seen that trend.
Stat trends seem to definitely be the best indicator though.
Like I said, there are exceptions. That one year for Mustard looks like an exception too. He has been terrible beyond that.
Poor speed influences and does so enough to negate good defensive ratings. I guess I should have worded it that way from the beginning.