The Effectively Wild Season Preview Series for 2015 is in the books, and with it came 30 predicted win totals, ruthlessly extracted from each podcast guest by Ben and Sam. Alongside it, Banished to the Pen produced its own series of season preview articles, many of which came with predictions of their own. For those that didn’t, I dragged a prediction out of the author in question, all in the name of putting together this prediction extravaganza. EW Facebook group member Tyler Baber also created a prediction poll that asked listeners whether they would take the over, under or agreed with the original PECOTA projection.

When I decided to gather all this data together, I was interested in answering three questions:

1. How optimistic (or pessimistic) were the EW guests and BttP writers; in other words, how far away from .500 was the cumulative league win total?

2. What would the divisional standings be with each version of the predictions?

3. How did the predictions compare to the collective opinion of the listeners?

I collated the predictions for all 30 teams from both EW preview episodes and our Team in a Box articles, and added up the responses from 153 EW listeners who responded to Tyler’s poll. In order to accurately present both the results of the poll and the changes made to PECOTA, I included both the PECOTA predictions from when the listener poll was created – February 6th – and the most recent ones, from March 31st. To add another level of context, I also added projected win totals from another source by using Fangraphs’ projected standings page. Finally, I put together a series of divisional tables below, along with some graphics to more clearly represent the distribution of the predictions. We’ll start by looking at the level of optimism in each set of predictions.

Level of Optimism

It seems almost inevitable that any uncoordinated set of projections from individuals such as this will produce more wins than could actually be possible, and so it proved once again. Effectively Wild season preview guests cumulatively predicted 72 wins over the actual 2430 that are possible, with 2502. That was actually an improvement on 2014, when EW guests topped the possible mark by 90. Banished to the Pen preview writers showed remarkable restraint to only predict 62 wins more than possible, even with a late 99-win prediction that ranked as the most optimistic individual win total of any included here.

18 of 30 predictions exceeded the most recent PECOTA totals, with four agreeing exactly. Of the eight that were under PECOTA, only two – for the Rays and Mariners – were more than three wins lower. Once again, the essay writers were far more likely to wildly exceed PECOTA’s projection than they were to undersell a team. Four predicted at least ten wins more than PECOTA and seven went more than five wins above.

BttP’s writers somehow managed to be both more and less optimistic. While 22 projections were actually higher than the most recent PECOTA output, none were more than 8 wins higher, and only six were over five. The most extreme difference was on the low side, as BttP’s Rays prediction was ten wins lower than PECOTA’s. Below is a clearer visual representation of how the four sets (BttP, the BP Essay writers, the more recent PECOTA and Fangraphs’ latest standings) compare to each other; it’s possible to remove any of the projections from the group if you’d prefer to just see the results for one, or compare two against each other.

Almost all of the extreme values are predictably from our human projections. Fangraphs is clearly shown here as having the narrowest range of values, while BttP writers just topped the EW guests for the widest range. While there was a good deal of agreement between the essay writers and BttP, including identical projections for the Angels and Mets, some particularly big differences also jumped out, especially in the cases of the Marlins and the Mariners.

Divisional Standings

A little extra optimism doesn’t stop us from putting together some projected divisional standings, even if the win totals don’t strictly make mathematical sense. Here’s how the individual divisions would look according to the guest predictions; just click the column headings to sort by alternative projections if you prefer. In all cases where percentages in parenthesis are included for the listener poll, I have cited the percentage of the option which got the highest number of votes.

AL East

What EW Listeners Said: The Rays’ original PECOTA projection was the most objectionable of any team for EW fans, as a whopping 135 – 88.2% – took the under on the 86 wins (now bumped up to 87), so perhaps that 77-win mark from BttP’s Brandon Lee would be much more to their tastes. That also lines up with our other projections, none of which exceeded a .500 mark for Tampa Bay. The crowd gave their wins to the Jays and Orioles, with 72.6% of listeners taking the over for each. Boston and the Yankees were the divisive teams here, as a majority could not agree on over or under for either, and 23 people thought that the 87-win prediction for the Red Sox was about right. All five teams were picked to win the division by at least six listeners: 79 went for the Red Sox, with Toronto second favourite at 48.

AL Central

What EW Listeners Said: The Tigers are far more favoured by EW fans on the whole, with 115 (75.2%) taking the over on 82 wins while Cleveland’s 81-win projection was deemed to be perhaps only slightly low, with 46.4% of the group predicting a higher total, against 37.9% taking the under. Kansas City (74.5%) and Chicago (56.9%) were both considered to be more likely to exceed their projection, while even a 70-win mark was thought to be too optimistic for Minnesota by more than half, suggesting that BttP’s 68-win prediction would be more to the liking of at least those 83 listeners. Detroit was named as division winner by 81 listeners, with Chicago and Cleveland in a dead heat at 30 votes each.

AL West

What EW Listeners Said: This is more of a three-team struggle in the collective opinion of the listeners, with the under taken firmly in the case of the Angels at 90 wins (58.2%) and marginally in the case of the Mariners at 87 (43.1%), with a slight lean to the over for the Athletics (41.8%). All three teams finishing with 85 or 86 wins would presumably not be a surprise to the Facebook group. They also expect a much bigger gap between those three and the Texas teams, with opinion clearly on the side of the Astros returning to the basement after the Rangers’ 2014 disaster (although this poll was taken before Yu Darvish had Tommy John surgery). Despite the quantity of listeners taking the under on Los Angeles, 85 still picked them for the division, comfortably ahead of Seattle’s 47 backers.

NL East

What EW Listeners Said: The listeners do like the Marlins but don’t like the Mets, with 60% taking the over on 81 wins for Miami and 51.6% also picking the Mets to finish .500 or below. Washington reaching 92 or above was considered unlikely by only 24.8%, while PECOTA’s 70-win Phillies projection was regarded as either accurate or too high by 85% of respondents. The Braves were also regarded as slightly more likely to finish under 73 wins, although over 20% agreed with that number. To underline how utterly predictable everyone thinks this division is, Washington received 149 votes for first place, with 3 going to the Mets and one voter (a disgruntled Phillies fan?) mysteriously refusing to pick.

NL Central

What EW Listeners Said: Many listeners may have found PECOTA’s Pirates projection more outlandish than Sprague’s, with around 80% taking the over. 89 wins was deemed around the right range for the Cardinals, with nearly a quarter – the highest total in the ‘agree’ bin – picking that number exactly and a  lean towards the under from the rest. Just over half of listeners thought 82 wins wasn’t enough for the Cubs, while the group feels Milwaukee (51%) and Cincinnati (62.7%) are heading firmly in the other direction. St. Louis received 109 votes for first place, with the 25 voters propelling the Cubs’ hype wagon into second place ahead of the 15 who like the Pirates.

NL West

What EW Listeners Said: 97 Dodger wins is a bit rich for EW listeners, with almost as many (128, 83.7%) objecting to that as the Rays’ projection. There’s a very slight edge given to the Giants over the Padres and the group is more inclined to knock wins off the Diamondbacks’ 74 (56.9%) than it is off the Rockies’ 72 (48.4%). Los Angeles is so far ahead that despite all the listeners taking the under, 137 still took them to win the division. Nine were won over by AJ Preller’s San Diego overhaul, while six laughed in the face of San Francisco’s apparent inability to win in odd-numbered years. Again, a voter abstained; Nora Morse, perhaps?

Listeners vs PECOTA

After all those percentages and over/unders, it seemed sensible to try to display the listener view of PECOTA in one clear graphic. The size of the bubbles in the below graph therefore represent the relative quantity of wins projected by that original PECOTA release, while the colors indicate the collective listener opinion, represented here by the total number of voters either agreeing with or taking the over on the projection.  A deep red bubble suggests that the vast majority of listeners took the under, while the bold green is an indication that they took the over. Here we can see the strong unders for the Dodgers and Rays, and the opposite for the Royals, Tigers, Blue Jays, Orioles and Pirates. The Yankees and the Rockies also show up (or rather, don’t) as the teams for which listeners couldn’t really decide either way.

Overall Standings

I have included all 30 teams here, as well as the EW listener over/under numbers. In addition to sorting by any of the options, you can filter by league or division by using the search box, and display more or less teams using the drop-down menu. The EW guest predictions would produce an all-AL East wildcard, with the Blue Jays at the Orioles, and San Diego heading to Pittsburgh in the NL matchup. They see a bigger distinction between the top and bottom in the NL than any other set of predictions, with both the bottom and top four coming from the NL (although the Angels tie the Nationals for fourth at 92 wins). The AL is conversely very bunched together, with a spread of just 15 wins across the entire league.

BttP’s preview writers have the Mariners narrowly missing out on a division title and instead hosting the Tigers in the wild card game. The Pirates again host the NL game, but this time it’s a repeat of 2014 as the Giants travel to Pittsburgh. Although there’s a more even distribution of the two leagues at the top and bottom than the podcast guests, BttP did compile the largest range of predictions, with a full 35 wins separating Matt Trueblood’s 64-win Phillies from Cline’s 99-win Nationals.

Composite Standings

One final experiment. Generally speaking, aggregating projections is more likely to produce more reliable results than any individual projection. To avoid the overzealous EW and BttP numbers having too much influence on the overall numbers, I have adjusted all the projections downwards to add up to the actual win total, whilst maintaining the same ranking by multiplying them all by the same number (2430/aggregate total). This is a rather quick and dirty method of getting to our endpoint, but then we were starting with sets of projections that were 60-70 wins too high in the first place. Below are the projected standings if you take an average of all four sets.

I’ll return to these at the end of the season to see whether they fared any better than any of the individual sets, both in terms of predicting win totals and the final order of the standings.

So, it’s clear that people are generally still more optimistic than projection systems, that the NL is a foregone conclusion while the AL is anybody’s (well, except the Twins), and that no team, not even the Phillies, will lose 100 games. I look forward to finding out exactly how wrong everyone was at the end of the season.

My thanks go to all of our excellent preview writers, who put up with me pestering them for predictions as well as doing a great job on the previews themselves. Thanks also to Tableau Public for the graphics, Fangraphs and Baseball Prospectus for the projections, and of course to Ben, Sam and all the EW guests for providing many hours of information and entertainment, as well as those crucial win totals.