Some Not So Fun Math
The Mariners are 13-21, 6 1/2 games out of first place. They have 128 games left to play.
Heading into the season, we had the team pegged as about an 83 to 85 win “true talent” team, which works out to about a .520 winning percentage. If the Mariners win 52 percent of their remaining games, they will finish with a record of 80-82. If our opinion of the team was exactly the same as it was on Opening Day, we would now expect the team to finish just below .500.
Of course, the last six weeks of baseball provide some evidence that there are reasons to think that this team might not be a true talent 83 to 85 win team. Milton Bradley didn’t stay under control. Jack Wilson didn’t stay healthy. Rob Johnson didn’t learn how to catch over the winter. Even leaving aside the slumps of guys like Chone Figgins and Jose Lopez, which have been big problems but should theoretically not continue for much longer, there are some structural issues with this roster that weren’t true in March. I’d probably argue that this is more like a 78 to 80 win true talent team right now, given how the roster has shaken out.
If I’m right, and the Mariners play like their “true” selves the rest of the year, they’d finish 75-87.
How wrong do we have to be for the Mariners to win the AL West? With 128 games to play, the team would have to play .601 baseball the rest of the year to finish at 90-72. Texas is good enough to win more than 89 games, and so I don’t think we can assume that anything less than 90 will get it done. .601 baseball is basically the level the Phillies have played for the first six weeks of the season. Right now, a .601 winning percentage would rank 6th highest in baseball. Even good teams have a hard time playing .600 baseball for four and a half months. And 90 wins is no guarantee of a playoff spot.
95 wins probably gets you in. It’s tough for me to see two AL West teams finishing with more than 95 wins, so we’ll say that’s the magic number to where you can be pretty darn sure you’ll be playing in October. The Mariners would have to play .640 baseball the rest of the year – about the level that the Yankees have played in the first six weeks of the season – to finish with 95 wins. That’s basically winning two out of every three games, all year long, starting today.
Like I said yesterday, I’m not trying to sound melodramatic, but the Mariners chances of making the playoffs now stand somewhere between five and 10 percent. They’re probably closer to five percent, honestly. It’s pretty bleak.
The point of this isn’t to be Debbie Downer, but rather, to suggest that the Mariners have probably played themselves out of a position to make the kinds of trades that people have been calling for. The numbers don’t add up to make a prospects-for-veteran swap make sense. Even if you could get Houston to trade Lance Berkman, this team has played themselves out of position to give up what it takes to get a guy like that. With trade season getting ready to open up, the Mariners can’t be “buyers” in the traditional sense. They shouldn’t make another Cliff Lee trade – there aren’t good enough odds of winning this year to justify it.
That doesn’t mean they shouldn’t make any moves and try to salvage the season. They aren’t so far out that you can pull the plug and just call it a rebuilding year. But they have to change the type of trades they may have been looking at. Instead of moving prospects for a big thumping bat who will walk away in a year or two, the team should be looking to make moves for guys that can help in 2010 and beyond – guys who make this roster better, but can also offer some longer term value.
Just as an example, the M’s should absolutely be burning up the phone lines in Detroit trying to get Ryan Raburn from the Tigers – a 29-year-old right-handed super utility player with a pretty solid bat and a good glove in the outfield who can also fill in at 2B/3B when needed. He makes the league minimum, is arb eligible for the first time this winter, and the Tigers just optioned him to Triple-A out of frustration with his slow start. ZIPS projects him for a .344 wOBA going forward, which would make him the second or third best hitter on the team, and he’d be a perfect complement to Saunders in LF while also offering a bat who can play multiple positions and have enough thump to improve the offense.
He’s not Lance Berkman, but he’s the kind of guy the Mariners should actively be trying to acquire. He won’t save this season, but there’s a pretty good chance that it’s not able to be saved, and the team doesn’t have a strong enough farm system to waste assets on a guy who won’t be here past this season.
They have approximately six weeks to turn this thing around and get back into the race, at which time they’ll have to figure out if they’re close enough to contention to keep Cliff Lee. They can’t just sit around and wait for guys to start hitting until then, but neither has this team earned the right to have the front office trade more future for present. That’s just not a wise trade anymore.
Whether they’re better than this or not, the 13-21 start has put the season on the brink.
Realistically, how much can we expect to get for Cliff Lee? Is he more valuable right now, or at the deadline? And will he be more valuable as a trade chip than the two picks we’ll get when he leaves?
I’m of the opinion that if Bedard comes back and pitches well, that the rotation without Lee can be good enough to be competitive. Unfortunately, our offense isn’t good enough to be competitive, so we need to trade a real piece of our current puzzle to get something to be excited about in the future. I’d rather be excited to watch some young guys develop than just happy about seeing a Cy Young winner show off at Safeco in his contract year.
This offense is epically bad
Figgins has a lifetime BABIP of .339. His current BABIP for the season is .241.
Jose Lopez has a lifetime BABIP of .283 and lifetime ISO of .137. His current BABIP for the season is .242, with an ISO of.065. Oh, and he has a history of going into huge slumps and then going on tears.
Essentially, you either have to argue that Figgins/Lopez are done as effective MLB hitters, or they’re in a slump they will eventually get out of. I’ll pick the latter.
Earl Weaver’s winning formula used to be pitching and three run homers.
Sure. That’s great when you have Frank Robinson, Cal Ripken or whoever in the 1970’s and 1980’s. Where do the cheap home run hitters come from in 2009?
Playing at Safeco, we should be looking for a LH power bat capable of filling the role that Branyan had last year. It doesn’t look like anyone currently on the roster has that kind of capability.
I liked Branyan as much as the next guy, but he has a lifetime wOBA of .350, and projecting him to repeat 2009 as a 35 year old with a bad back isn’t reasonable. Michael Saunders is capable of developing into that level of hitter (though his skills will be different), given his minor league numbers, and Milton Bradley’s wOBA last year, during a DOWN year, was .345.
There are problems with the roster (C and DH are huge piles of suck, SS is one too if Jack Wilson isn’t playing regularly, Kotchman is looking more and more like a more expensive version of Mike Carp with a worse bat and better glove), but the answer isn’t going crazy over power and ignoring everything else. Adding better talent at SS and C and installing Bradley as the everyday DH and Saunders as the everyday LF would go a long way towards making the M’s better, and wouldn’t necessarily involve adding a HR threat.
The Dallas Morning News says that the Rangers would consider dealing Smoak for a player that will put them over the top. Despite the fact we are in the same division, that would be an exchange that could make sense for both teams.
Since we are falling out of the race much quicker than expected, Jack Z will have plenty of time to field and compare offers for Lee and whoever else is in demand.
Oh I wouldn’t go crazy over power. I would settle for one legitimate HR threat. The Mariner power currently is like trying to light up a big room with a 30W light bulb.
I expect Figgins and Lopez to revert closer to career norms. That won’t be sufficient to turn the team into a serious contender. It would help them move towards .500, though.
Bradley has the potential to be the kind of power threat the team needs but, in addition to his emotional problems, his skills may be in decline according to some scouting reports.
I am hoping Z can find some power magic similar to Jack Cust. How about trying to get Kila Ka’aihue from KC? He is blocked there and could be a good fit at Safeco (don’t know much about his defense, however).
The problem trading with Texas (ie Smoak) is that I’m not sure MLB will let them make trades at this point, since they’re considering taking over the day to day operations until a sale goes through. Not to mention if Texas is doing anything, they’ll be going after Roy Oswalt who just announced he’ll accept a trade as long as it’s to a contender, and has mentioned Texas as somewhere he’d be willing to go– but again, the Rangers are not in a great position right now.
Unfortunately (regarding Smoak) but fortunately (regarding Oswalt).
I know folks are tired of Adrian Gonzalez speculation, but I still can’t help but wonder about Cliff Lee being an attractive enticement if San Diego is still in the playoff picture and we’re NOT come July. If either Kotchman or Lopez can heat up between now and then, it could be even more interesting to SD.
If they can’t afford to keep Gonzalez, he’s basically a rent a player, just like Lee, and a dominant pitcher is a great commodity come October.
Dave, I’ve seen you give similar figures in the past but in a sense isn’t something like this puting undue weight in a small sample? To expand if we think someone is a true talent, BABIP luck aside, .333 hitter, isn’t that our guess for the whole season, like in a projection system?
I mean if I project someone to hit .333 I know very well he isn’t going to come up and get a hit literally once out of every three times. He might hit .450 for all of April but based on the knowledge I used to make my guess I now in May am going to expect a slump to come, maybe he hits .220 for this month after a prolonged o’fer. Or maybe the slump doesn’t come till July.
Now I realize your math is right on that in order to contend we’re going to need to play ~.600 ball the rest of the year. But if one thought they would win 90 games to start the year they probably didn’t think they would win exactly 15 games each month, there most likely would be some kind of month to month variance. So with that in mind if one still thinks this is a true 90 win talent team couldn’t they manage to play .620 ball the rest of the year to cancel out the bad start, have a couple prolonged winning streaks to make up for the big losing ones?
Like flipping a coin 1,000 times you’ll more than likely get somewhere around 500/500 but maybe the first 100 went 65/35? Or is a baseball season not long enough to allow for those things to normalize? Maybe there is an article you can point me to that shows how the luck factors involved in determining who scored more runs on a given day make it so once a game is in the books it shouldn’t be factored in with the other games in a season and allow a winning percentage rate to regress to the mean.
Sorry to kind of ramble but to summarize it seems like you treat overall team record for the year totally different than a pitcher’s HR/FB rate (or whatever stat) where you expect that stat to normalize by the time the year is over, but the wins once in the book you seem to think should no longer be looked at as a whole with the wins going forward and only expect the team to perform at the preseason expectation level going forward. Is there a specific reason for this differentiation? Thank you.
That’s the Gambler’s Falacy ec linked to above. A string of bad luck doesn’t imply that you’re due an equivalent string of good luck. You get what you get. Luck has no memory.
A 5% chance! That sounds pretty good. I’ll put all my money on that…
If the Padres are in contention as the trade deadline approaches, do you really think they are going to trade A-gone? The Padres have a club option for 2011 for $5.5MM. That is a bargain, even if he only produces 3 WAR. I would hardly think he is a “rent-a-player” given they have him for potentially one more season.
Sometimes reality is a hard concept to grasp.
Hopmacker–
“Sometimes reality is a hard concept to grasp.”
A little harsh as it was intended, but I agree. Cliff Lee was a combined 13.8 WAR player the past 2 years. Lopez was predicted to be 2 WAR this year.
So you wouldn’t trade a 3 WAR 1B for a 6-7 WAR pitcher and a 2 WAR (Lopez) 3B/1B? (Padres also need depth at infield).
And lest we forget, Lee was THE best pitcher in all of baseball in October last year.
I guess it depends on how you view offense versus pitching, and I’m counting on the Padres to be true to their NL roots.
For a team as offensively starved as the Mariners are, I can see where you don’t smell anything in that deal, but San Diego may be different if they’re starting pitching isn’t able to maintain.
(Hell, if the M’s are playing bad enough, we might even be able to swing a waiver deal later in the season.)
On the other hand, a win streak changes the to-the-end-of-year calculations. If the Mariners were to go 10-2 over the next 12 games (playing .500 ball over 24 games), they then need to play .580 ball the rest of the way to get to 90 wins.
A win streak does wonders for the W/L. It probably doesn’t even matter where it is, as long as the players keep the faith. The Athletics did it in 2002 with a very late 20-game win streak.
If the Mariners played .580 ball for a month, then had a 10-2 stretch, then they are on pace for 90 wins by finishing with .580 ball.
If it would happen is another question.
Unfortunately no, that’s not how statistics and probability work. Many people think that that is how “the law of averages” works, but the law of averages does NOT work that way.
Now it is true that if you’re predicting the season as a whole, and you predict that the M’s will win say 55% of their games, then you can also very safely predict that somewhere in 2010, the Ms will have one or more cold streaks, counterbalanced with one or more hot streaks, such that overall the most likely outcome is 55% wins in the end.
But that’s not the situation we’re currently in. We’re not predicting the season as a whole, because over a quarter of it has already gone by. Our task is to predict what percent of games the M’s will win the rest of the season. If we initially believed that the Ms would win 55% of their games, and if they have so far won 38% of their games, we have three choices:
a. Hold fast to our pre-season estimates, and believe that the early season losses are due to random error, and therefore the Ms will win 55% of their remaining games.
b. Use a Bayesian approach, and modify our estimates based on the observed data. Our pre-season theory said 55%. A small-but-still-substantial set of data says 38%. Our new, modified estimate will be somewhere in between, with the relative weight dependent on how confident we are of our pre-season estimate, and how large the empirical sample is. But the resulting estimate will say that the Ms will win <= 55% of their remaining games.
c. Predict, fallaciously, that the "law of averages" will have to take hold and the Ms will magically counterbalance their cold start with a hotter-than-55% hot streak, and thus the Ms will win more than 55% of their remaining games.
Unfortunately, there is no (legitimate) theory of statistics or probability that chooses (c). It either (a) or (b).
Now it is true that the Ms might get lucky and win more than 55% of their remaining games. There's always a chance for luck. That's why we play the games, etc.
But there is no theory which permits us to *expect* that the Ms will magically bounce back and win a higher percentage of games that our pre-season estimates. It is a possibility, but it requires a lot of unlikely events to take place.
The Law of Averages does permit us to legitimately expect that Chone Figgins' batting will improve, Lopez's batting will improve, etc. But "improve" means "return to their pre-season expected levels" not "bounce higher than their pre-season expected levels".
I guess what I was getting at was option A there MKT. It seems that when a pitcher has a really hot start but a 0% HR/FB rate for example most Fangraphs authors or the like blame that on random error and expect it not to continue. So of course over some sample going forward they would be post an above average rate and in the end wind up with an averagish rate. But with wins losses it seems they look only at the option B you present, I was just wondering if there is some reason option A is completely dismissed with wins losses.
Option A is exactly what Dave was using in the second paragraph of his posting … Option B is what he used in his third and fourth paragraph. In his fifth he used Option C or more accurately, he pointed out that in order for us to believe that the Ms will contend, and barring some miraculous transactions by Zduriencik, we’d have to believe in Option C. And Option C is basically the same one that Royals fans and Pirates fans use each April if they hope that their teams will contend: pray for a miracle.
That is not correct. His winning forumla was pitching, defense and three run homeruns, but he has been on the record for saying that he only used that strategy because it was the best use of the enviroment. He noted in his book that for teams like the Royals (who played in a giant pitcher’s park back then) going for speed instead of power is a good move and that his team always had problems playing at KC because of that.