Not sure if this is the right spot but when looking at MLB games what stats are most important. I think looking at pitchers I do era, k's per inning, whip and record. Team hits, obp hr and then look at the last 10 days for the team. What do you think is most important?
-
-
I don't think you can really handicap MLB just looking at a few stats. There are two different duels going on: visitor hitting versus home pitching, and vise-versa.
Every approach I have seen that works evaluates each team's hitting, starting pitching and bullpen pitching.Comment -
And, looking at just the last 10 days in MLB is typically a recipe for failure.Comment -
I'm considering so different than other bettors. I don't care W-L record (against todays rival or general) for pitcher which is the most nonsense stat in baseball. just curious why people care...
also I'dont care so much ERA.there are numbers of factors. Actually the most important thing is watching previous matches but if we are talking about stats;
*are batters hot? what are the AVG and OPS numbers on last matches? and against who? maybe 0,217 against halladay is not so bad but 0,222 against R.Lopez+Samardzija is not so good.
*whip (but against who? maybe he played against against redsox,rangers and yankee on his last 3 matches ?)
*Starting Pitcher P/IP on last 5 matches (average pitch number for each inning) for example Karstens shows wonderful performance on last matches. also could he make S. Out? how many times? against who?
*what about starting pitchers ground ball / fly(air)ball odd? if it is over 1,4- 1,5 that is good. if his rival has hot and powerful hitters under 1,00 is bit of dangerous. (for example : Ogando Vs Yankees)
maybe I have a bit of strange english
I'm sorry
Comment -
^^that's pretty much a good basic strategy. Already ahead of most. Justin7 makes some good points as well. As a basic rule in baseball today a good strategy is to find out what players' numbers are quite exaggerated and use that to your advantage. If you don't think a pitcher's whip can be 1.13 but he's really running well-above expectation and should be at 2 then you are mistaking. People don't realize the power of variance and how it rears its ugly/beautiful head deep into the season.Comment -
If you are calling baseball games "matches," and/or you are claiming that recent performance is really important, chances are....you're not winning.Originally posted by gangeriverI'm considering so different than other bettors. I don't care W-L record (against todays rival or general) for pitcher which is the most nonsense stat in baseball. just curious why people care... also I'dont care so much ERA.there are numbers of factors. Actually the most important thing is watching previous matches but if we are talking about stats; *are batters hot? what are the AVG and OPS numbers on last matches? and against who? maybe 0,217 against halladay is not so bad but 0,222 against R.Lopez+Samardzija is not so good. *whip (but against who? maybe he played against against redsox,rangers and yankee on his last 3 matches ?) *Starting Pitcher P/IP on last 5 matches (average pitch number for each inning) for example Karstens shows wonderful performance on last matches. also could he make S. Out? how many times? against who? *what about starting pitchers ground ball / fly(air)ball odd? if it is over 1,4- 1,5 that is good. if his rival has hot and powerful hitters under 1,00 is bit of dangerous. (for example : Ogando Vs Yankees) maybe I have a bit of strange english I'm sorryComment -
Recent performance CAN be important...but for the opposite reason that people think.Originally posted by evo34
Bad players that are running hot and good players that are running cold in recent games generally tend to regress to their norms, so fading the recency can be good.Comment -
What about night era and day era if it is night or day (obviously) as well as road/home avg of opponents then various yet similar bullpen stats as well as batting stats? I'd obviously have more stats but would you consider those important? Just an idea.....Originally posted by Justin7
And thanks to everyoneComment -
^^^absolutely useless. Shit like they makes me laugh(not a dig at you btw. People who have done this for years still cite dumbass statistics as their handicapping reasoning.). Average vs. Pitcher. Home average vs away average. Day game vs night game. Playing west coast teams. In interleague. 1st game back from road trip. Following a full moon loss. All of the incomplete sentences I just listed are fukking pointless stats that people think matter.Comment -
Some home/away, day/night splits have statistical significance.Comment -
^^^no they don't. They have the illusion of statistical significance because people don't understand statistics.Comment -
illfuuptn,Originally posted by illfuuptn
Are you prepared to back that up?Comment -
interestingly, toronto is not known as being a homerun park but the blue jays destory the ball there considerably more than opponents. there are rumors about them stealing signs, which is what giradi was talking about several days ago.
as far as baseball stats. you should go to fangraphs and read as much as possible and learn about the objective analysis of baseball to get ideas for statical analysis. for example, this year the rays are one of the best defensive teams in the history of baseball. they really take a lot of pressure off the pitching staff. how much that attributes to wins is debatable considering that the phillies and yankees are two of the worst defensive teams in baseball up the middle(where defense is most important) but are two of the best clubs.Comment -
Originally posted by illfuuptn
100% guarantee you won't be able to back this up with anything relevant. This will be good.Comment -
oh and fangraphs also updates projections for the rest of the season if people are using the methods talked about in justin7's book. he stated in the book the edge was lost after a few months and wasn't able to find an adequate way of making changes to preseason stats. i haven't tried using the updated projections to know if they are viable, but it is something to consider.Comment -
If you have X number of baseball players, it's not surprising that you have 1/20*X outliers, or something like thatOriginally posted by tim0402Comment -
Originally posted by antifoil
do they have team projections or just player?Comment -
I would have to disagree to this. To say those stats are "absolutely useless" is both baseless and ignorant. I wonder what type of statistics you do deem important if you are completely discarding these type of stats in your handicapping. Human beings play sports and as such, I would argue that these type of statistics are more relevant than those you use. Day/Night statistics pitcher? I think it matters. Some pitchers have such a huge disparity that it is somewhat foolhardy to even attempt to argue that it is "absolutely useless". Think about how YOU, yourself would perform under such a circumstance? Perhaps, it will have no effect to you personally whether you were to pitch at day or at night. But, is it so hard to believe that it may have a great influence on someone else? And, how could you even say that average vs pitcher stat is "absolutely useless"? That is just comical. These are human pitchers who throw different stuff. Not a ball machine that throws the same stuff over and over. Hitters like to hit off certain pitchers. Maybe, you've never played baseball in your life. Yes, that would explain it.Originally posted by illfuuptnComment -
Just because most are absolutely useless does not mean they all are. See for example, Josh Hamilton's day game numbersOriginally posted by illfuuptn
The key for these is two-fold: predictive power and market ignorance. Plenty of them might actually have some predictive power but if the market is aware and prices it in it's useless
If you are trying to predictively model games these are all useless. Try SIERA, K/PA, and instead of WHIP regress BABIP to the league average(for the majority of pitchers) or their historical averages for the few exceptions that have proven to be sustainable(cain, weaver, etc)I think looking at pitchers I do era, k's per inning, whip and record.Comment -
Most abnormal splits (day/night, certain lineup position, player in a certain stadium) are in fact bullshit, aka random, aka non-predictive. Not that I am trying to convince anyone not to chase them... Opportunities for edge don't self-generate -- they are created by others. Whether it's people chasing W-L records, hot streaks, small-sample batter vs. pitcher stats, whatever. The only major exception is platoon splits. But people don't model them -- esp. for pitchers -- nearly well enough.Originally posted by tim0402Comment -
Well, there is no need to actively fade anything. Just have a model that doesn't do anything special with recent performance.Originally posted by LT ProfitsComment -
you can multiply whip * pi for each player and get an approximation for the entire league average era, since its expressed in ERA form its useful for modeling
the new SIERA stat at Fangraphs, along with xFIP, is the best measure of pitcher performance IMO
you can try to isolate tendencies with LD% and BABIP then project to ERA through a regression analysis or simple arithmeticComment -
This, imho is totally wrong, and I'm not talking about the fading nonsense.Originally posted by evo34Comment -
Go for it. Tell me the proper weighting of very recent performance vs. long-term performance when predicting a player's skill level for the current game?Originally posted by chunkComment -
First, I prefer to deal with this issue in the context of team performance as opposed to individual, but it is a valid question. Are you looking for a variable to plug into a formula? For now, I am going to answer your question with a question. In the course of a 162 game schedule, would you give more weight to the performance level of their 1st 81 games or their last 81 games? I know that this is generic and simplified for illustration, but the answer is obvious in my opinion.Originally posted by evo34Comment -
I am not really talking about auto-fading, I am talking about taking advantage of market overreaction to recent performance, so it is more of an indirect "fade".Originally posted by evo34Comment -
some pitchers are better @ home, there is no disputing that.Comment -
Have been better at home, or will be better at home? Two very different things. Also, average HFA and park factors come into play, so being "better" at home means relative to the expected home performance.Originally posted by tim0402Comment -
Assuming no personnel changes (which is an obvious pre-requisite), I would weight them very similarly. Slightly more weight to the second half, but not enough to make it important to do so.Originally posted by chunk
Of course, the OP mentioned looking at last 10 games. I suggested ignoring recent performance anomalies. You said that was "totally wrong." Still waiting for any evidence I am totally wrong.Comment -
like I said I'm sorry for strangle english.that is not my native language.I'm an europan and people call as "match" while talking about sport meetings.Originally posted by evo34
soccer,basketball, cricket or boxing.that is an habit and even I don't care. it seems americans call as "game". there are a lot of different calls.
you call "football" we call it as "american football" , you call soccer, we are calling it as "football" , americans say racetrack, but it is "racecourse" in UK english.
I understand that if I speak UK english I can't win on MLB. Thanks
Comment -
Originally posted by gangeriverlike I said I'm sorry for strangle english.that is not my native language.I'm an europan and people call as "match" while talking about sport meetings.
soccer,basketball, cricket or boxing.that is an habit and even I don't care. it seems americans call as "game". there are a lot of different calls.
you call "football" we call it as "american football" , you call soccer, we are calling it as "football" , americans say racetrack, but it is "racecourse" in UK english.
I understand that if I speak UK english I can't win on MLB. Thanks
do you call bettors/gamblers, punters?Comment -
it's a team game so dont look at the stats of one single player, for example the starting pitcher.
i also want to add that pitcher's era should not be the basis for o/u.
GLComment -
^^ And of course you don't only look at starting pitcher you would need to look at bullpen. Like for example if Braves go Jair for 7 innings then Jonny Venters and/or Eric O'Flaherty to set up Craig Kimbrel they wont give up a lot of runs but if they have sucky set up and closers it's completely different.. those pitchers are equally important to starters
And has anybody done research on errors correlating with runs?Comment -
punter or bettor.Originally posted by tim0402
gambler looks like a person who is out of betting.he/she is maybe a poker player or who interested in slots, I mean a casino player.Comment -
Better wording would have been wrong to totally ignore. I agree with LT Profit in that recent play can be an important part of the picture. The quest is never ending and at the end of the day if it is all just random, maybe we can all have fun deluding ourselves.Originally posted by evo34Comment
