How good (and lucky) do the Braves need to be? (updated)

by Bronn on Aug 12, 2009 7:11 PM EDT

As of this post (prior to the 8/12 game against the Nationals) the Braves are sitting at 59-54, 4.5 games out of the NL East lead, and 3 games out of the wild card lead. The Braves' recent surge, which is highlighted by a near-sweep of the Dodgers, indicates that they may be better right now than their record indicates. but just how much better are they?

Going back to June 28, the Braves have gone 25-14-that's a .641 winning %. As I've stated in an earlier post, the likely number of Braves' wins needed to reach the postseason is 90. That's based on several projection models, but mostly can be seen here and here. I won't go into those numbers any further at this point. Obviously this isn't concrete, as the Braves themselves can play a significant role with their remaining games against the Phillies and Marlins, but I'm going to continue assuming that 90 wins is the desired goal.

With 49 games remaining, continuing to win 64.1% of their games puts the Braves exactly on the projected 90 win pace. So the question is, are the Braves able to continue winning 64.1% of their games, given that they've spent so much of the season hovering around .500? The fact is, that over just 39 games, a .500 team has a 23% chance of maintaining a .641 winning %. So how do we know this isn't random variance?

This is where you have to add things that you know to a probabiliy picture. At the beginning of June, the Braves replaced a largely-ineffective Jordan Schafer (1.62 WARP over 162) with Nate McLouth (4.8 WARP over 162), and then replaced a just-as-ineffective Jo-Jo Reyes (-3.6 WARP over 30 starts) with Tommy Hanson (6.5 WARP over 30 starts). If you assume the Braves were a .500 team prior to these moves, then this transforms them into a team expected to win about 94 games over a full season. Unfortunately, that translates to a .58 winning %, when Atlanta still needs to win 64% of its games.

Additionally, these two moves did not pay immediate dividends-the Braves had a horrible month of June. This can be largely attributed to the tough schedule-including 6 games against Boston and 3 against the Yankees, but the struggles of Derek Lowe and, to an extent, Jair Jurrjens are also evident.

This June Swoon can be largely attributed to random variance, as the Braves rebounded somewhat at the beginning of July, winning 6 of 10, including a sweep of the Phillies. But then FW pulled off his masterstroke move, swapping Jeff Francoeur (-1.58 WARP over 162) for Ryan Church (5.67 WARP over 162). This move alone could turn an average team into a contender. Since this trade, the Braves are 18-9 (that's .667 winning %, if the math isn't too easy for you). If the Braves had been a .500 team prior to June 5, they would now be a team expected to win about 101 games over a full season. (In case you haven't noticed, I'm using 162 for all the players I'm evaluating, which doesn't factor in injuries or random days off, and gives full credit to guys who are being platooned.) The Kotchman/LaRoche trade I'm leaving out, as I'm looking at WARP added to the Braves, and the sample size is just too tiny for that move right now.

So, were the Braves a .500 team before all this happened? Going back to June 4, their record was 26-26. This is just prior to Nate McLouth's first game in a Braves' uniform, and likewise not long before Tommy Hanson's promotion. I'd like to accept the record for truth at that point, but since I'm not being lazy, I can just use pythagorean W/L (we'll pretend that's the gospel, even though it's not perfect either, okay?) Fortunately, the Braves had 228 RS and 232 RA on June 4, which gives them an expected W/L of .500 as of June 4.

So, based on the fact that the Braves were, in fact, a .500 team on June 4, they have made enough moves to, according to WARP, turn them into a 101 win team. Is that enough? A 101 win wins 62.3% of its games. That means, the Braves, if all plays out accordingly, should be expected to win 30 of their remaining games, putting them at 89 wins. So, with just a tiny bit of luck, the Braves should be able to reach 90 wins, if that is indeed the mark required for a poststeason birth.

Without going into the records of the remaining opposition, I'll quickly evaluate the Braves luck factor. For the value of luck, I'll use the LUCK Factor formula laid out by Greg Maddux, for which Luck = (8 games left against the Nationals) + (9 games left against the Mets) = Yeah...we're looking good.

This FanPost does not express the views or opinions of Talking Chop.

Tweet Comment 14 comments | 4 recs |

Do you like this story?

Comments

Nice post!! I honestly think that we are looking pretty good for the stretch run, but I think that we will need more than 90 to make the postseason!!

by macc on Aug 12, 2009 10:32 PM EDT reply actions

90 is still the number most project a few days later.

Remember, kids...don't ever let facts get in the way of your argument.

by MichaelProcton on Aug 17, 2009 11:06 AM EDT up reply actions

if we need more than 90 then honestly i think we are out,
but i think the phillies and the marlins will nose dive here shortly

by carolinabrave89 on Aug 12, 2009 11:14 PM EDT reply actions

but yes great post

very entertaining

by carolinabrave89 on Aug 12, 2009 11:15 PM EDT up reply actions

Not a fan of WARP. Kind of outdated, but at least they fixed some of the previous problems like excessively low replacement level and positional adjustments based on average hitting for the position. The batting portion is okay and will get better once the positional adjustments get completely fixed, but FRAA is just bad. “The RBI of fielding metrics” is a good description for it.

For pitchers, it’s based on some combination of runs allowed (the actual stat itself) and defense adjusted for team. BP isn’t clear on what defense is so it’d have to be something like a team-wide FRAA or defensive efficiency. These issues with pitching WARP inflate the value of pitchers with a low total of runs allowed despite shaky peripherals. Jarrod Washburn is a good example. Similarly, Hanson benefits from it since his FIP is above 4 and his strand rate is over 80%+ while Reyes is hurt by the high runs allowed despite a 5ish FIP.

Because of these issues, I prefer the WAR methodology. Better positional adjustments and defensive metrics and pitchers are valued based on metrics that filter out defense. I think Rally’s, Justin’s, Fangraph’s, or Tango’s version of WAR would do a better job than BP’s WARP.

I believe you took each players performance in a Braves uniform for this season and simply extrapolated it out to a full season or 30 starts. It’s good to look at what actually happened (man FYF was awful), but it does not necessarily reflect what would likely happen over a full season. Church’s value gets inflated here to near MVP-level because he’s on a hot streak right. I like Church a lot and think he can be an average player at worst (when healthy) to an above average player, but not a player 5-6 wins above replacement.

I think WAR levels would be more accurate and look something like this over a full season.
Schafer: 0 to 2 WAR depending on health.
McLouth: 3-4.5 WAR (about +12 to 25 runs with bat, +22 runs rep level, +2.5 runs positional, -5 to -15 runs glove)
Reyes: 0 to 1 WAR
Hanson: 2.5 to 3 WAR (assuming about a 4 FIP over 170-180 innings)
Francoeur: -1 to 0 WAR
Church: 2 to 3 WAR (about +0 to 10 runs with the bat, +22 runs rep level, -7.5 runs positional, +5 to 10 runs glove)

I made the run totals for hitting and fielding kind of broad because those players’ true skill level is debatable and I’ve already take up enough space. Take the average WAR of the players and our new players are about 8 wins better which sounds more reasonable. Add that to a .500 team’s record and you get 89 wins. There’s other factors too, of course, and we’re both simplifying by using .500, not factoring in the emergence of Prado, FUGA regressing from horrible to a solid hitter, etc.

Anyway, it’s obvious all 3 moves helped the Braves significantly. As of now, I think the true talent level of the team is above 90 wins and there’s a legitimate shot to make the playoffs. Upcoming Phillies series will be real intense.

by VictorW on Aug 12, 2009 11:21 PM EDT reply actions 2 recs

schafer was negative WAR, i know if healthy he would have been better but huge upgrade. Reyes another negative, replaced by a positive Hanson. Same with Jeff and Church

by McCann's the Man on Aug 12, 2009 11:23 PM EDT up reply actions

Reyes was actually produced positive WAR in the majors this year. A whopping 0.1 over 27 innings. I probably should have been more clear but those WAR numbers are sort of my projection for a hypothetical 2009 if they played the whole season based on this year’s performance and past performance. I think the WAR range is pretty solid depending on how optimistic or pessimistic you are about each player.

Anyway, if you want to take the low end of the replaced players compared to the average of the new players (I think I was a bit generous on McLouth), you end up with +10 wins or a 91 win team.

by VictorW on Aug 12, 2009 11:54 PM EDT up reply actions

WARP was kind of the lazy way out for me

I went with it, even though I know they have a really sucky way of calculating defensive value. But they also don’t exactly share their formula, and for the most part, what I see with WARP is generally what I expect to see, which is why I can find it useful. I don’t really have a problem where they evaluate “replacement level” as long as it’s uniform.

I did pretty much what you described, taking the performances in a Braves’ uni and extrapolating out over a full season. In the case of Church, it’s a bit inflated, but I also felt obligated to ignore the large part of his Mets’ numbers due to their home ballpark and his extreme home/away splits there. I didn’t have a decent way of giving Church a fair shake-he’s also completely healthy for the first time since 2007, and it didn’t seem right to use his partials for this year or to pull his 2007 numbers out for predictive purposes. But if the Diaz/Church platoon resumes once McLouth gets back, both guys will post really stellar numbers.

While my method has its failings, I still felt I got a very decent overall picture out of it.

One thing about Fangraph’s WAR is that I understand it even less than I do WARP. Not having read the book, I do know what wOBA is designed to do, and wRAA is supposed to be, but I have no idea how these are calculated. I went with comfort, in this case. And I avoided having to guesstimate where a guy’s true value lies. Likewise, while FRAA sucks, UZR isn’t that great over sample sizes less than full season anyway (and really, more than that), so it’s a bit imperfect to use for WAR.

I do appreciate the criticism, and I’m glad you were easily able to pick up on how I arrived where I did. While it might not be completely correct to describe this as a 101-win team, I feel like that can’t be too far off since we’re getting value from pretty much every position right now (including Garrett Anderson). The record since we got Francoeur out of here is phenomenal. Our starting 5 have all been great of late, and we have two great guys going in the back of the bullpen. It is a fairly marginal defensive team, though.

by Bronn on Aug 12, 2009 11:53 PM EDT up reply actions

The huuuge jump in wins from 3 players really stuck out for me. BP reworked WARP this year so it’s replacement is where it should be and, as a result, it’s more in line with WAR now, though it still has issues. For WAR, 2.2 wins above replacement is average, typically 4+ wins for an All Star, at least 6-7+ for a typical MVP (Pujols is nearly 8 wins a season), 5-6+ for a typical Cy Young.

Anyway, like you said, it gets the overall picture of the Braves right. Dumped some unproductive players for above average to good players while FUGA, Odarp, and LaRoche are hitting well too. Team is playing sooo much better than a month ago and it’s nice to see the ridiculously good pitching actually get run support.

by VictorW on Aug 13, 2009 12:27 AM EDT up reply actions

And if doesn't seem right to call this a 101-win team

Keep in mind that if you have have a 101-win team, on paper, in which every guy is playing 162 teams, that team won’t end up winning 101 games. There are always injuries, routine days off, and unforeseen declines by top-level relievers.

So, in short, I’d agree that this team, as currently constructed, is probably 95-98 win team over a full year. The big question now is whether they can get to 90. Given some fortunate scheduling, I think it’s entirely possible.

by Bronn on Aug 13, 2009 5:01 PM EDT up reply actions

Even 101 wins for a team on paper is pretty steep. But, more specifically, +20 wins from replacing 3 players stuck out at me since it’s the equivalent of replacing 3 replacement level players with 3 MVP contenders.

95ish sounds pretty solid, maybe a tiny bit optimistic, but not unreasonable. The team really has no holes. I guess the defense could be better, but every position can hit right now, the worst starter is league average, and the bullpen has at least 2 aces. Some of the playoff odds had the team at 25-35% which isn’t bad at all.

For simplicity, if we assume 90 wins is the cutoff for the playoffs, the team needs to win 30 of the remaining 48 games and finish as a .625 club. A 95 win team is a .586 team so I wonder what the odds are of a .586 team winning at least 30 games out of 48. Maybe I should have paid more attention when I took Probability and Statistics, but there has to be a fairly simple way to calculate some crude playoff odds to compare to the 25-35%.

by VictorW on Aug 14, 2009 12:28 AM EDT up reply actions

I think we cannot afford to lose a series to the Marlins or especially Phillies, and I think we will need to sweep at least one series from the Phillies in order for us to win the division. I also feel that as of right now, winning the division will be the best show we will have at making the playoffs.

by pancanbra on Aug 13, 2009 12:33 PM EDT reply actions

Nice post. I agree that around 90 wins is a minimum to win a playoff spot, that the Braves can reach that level, but they’d have to play amazingly well and get a good amount of luck to do so. But even if they fall short, it’s a lot more fun watching them play well than flounder like they have the last few years.

by redwards95 on Aug 14, 2009 10:41 AM EDT reply actions