Tuesday, April 19, 2016

Jose Bautista's Gambit: The Game Theory behind the failed takeout slide



Over two weeks after Blue Jays slugger Jose Bautista became one of the first victims of MLB's new takeout slide rule, I can't get over the fact that Joey Bats actually made what was a smart decision.

The slide itself seemed borderline. This wasn't the traditional spikes-up M.Bison from Street Fighter Slide Kick Into the 2nd Baseman takeout slide that Chase Utley used to injure Ruben Tejada in the 2015 playoffs. Bautista's slide was in line with the bag. His hand made contact with Logan Forsythe's leg as Forsythe tried to turn the potential game ending double play. The booth umpire decided that Bautista had reached for Forsythe's leg on the slide, and called the batter out due to runner interference. Game over, Rays win.

Bautista's no dummy. He knows as well as anyone that interference with the middle infielder on slides into 2nd is now illegal. And I don't believe for a second that his outstretched left hand was circumstantial. Every player knows to keep his upper body compact to maximize speed when sliding into a base. The only reason Bautista happened to extend the hand on the same side as the 2nd baseman was to interfere with his throw.

While that seems rather dumb, to knowingly interfere when it's not legal... Bautista's play was a great game theory move, and the best play for his situation.

------------

See, if Bautista slides normally, Forsythe, a competent 2B, turns the easy double play to throw out the not so fleet footed Jays batter Edwin Encarnacion at 1B to end the game. Of course, if Bautista blatantly slides into Forsythe, it's not only obvious interference and the game is over... but he could also be suspended.

However, when Bautista slides normally into the bag and subtlely extends his hand, there is a chance that the umpires overlook his interference, with the more substantial chance that Forsythe doesn't complete his play (which is what happened: Forsythe in fact made a throwing error due to interference from the outstretched hand), tying the game (Josh Donaldson would score from 3B) and keeping the game alive with two outs in the top 9th. Having a runner on 1st with two outs, obviously, gives you a better chance of winning than the game being over.

Yes, Bautista is often cited for interference and the game is over anyway. But it took an astute challenge from Tampa Bay manager Kevin Cash and a conclusive booth review for that ruling to occur. Bautista's interference was subtle, and there appeared such a substantial chance that he could get away with it that many argued (incorrectly) that Bautista hadn't intended to interfere at all! The umpires on the field in fact had not ruled he done so.

Of course, the opposing manager and the booth were a little sharper than that, and Bautista's gambit did not pay off. It also sent a message that this subtle hand-checking-like attempt at interference will get called, so it's unlikely players will attempt it in the future.

But before we knew all that, Bautista made the game theory optimal decision that it was worth a try. It wasn't clean, and he got caught, but it was a very smart move on his part... smarter than playing it clean and taking the certain defeat. There was greater expected value for him and the Jays in attempting to get away with interference, than there was in sliding normally... even with the high odds that his gambit would not succeed.

Tuesday, March 15, 2016

Breaking Loose in Tulsa: Anatomy of a bad at-large NCAA Tournament selection, and the business of at-large bids

Every NCAA Tournament selection show ends with howling over snubs and undeserving inclusions. This year's bad-inclusion winner is a duesy: Tulsa made the field at large despite ranking 75th in Massey Ratings and a mediocre finish that included multiple losses to unheralded Memphis.

Syracuse (58th) caught similar flak, especially after losing 5 of their last 6, including a regular season closing road loss to NIT-worthy Florida State. This is especially glaring given the omissions of worthy programs like South Carolina (44th) and Georgia Tech (45th). However, Tulsa's inclusion is quite glaring given their rating even made them a marginal pick for the subordinate NIT tournament.

Someone on reddit floated an excellent theory that might not only explain these inclusions but also illustrate the sort of political factors within the NCAA that lead to certain selections and omissions:

Each conference gets a certain sum of money for every NCAA bid they get. Louisville and SMU, who would have certainly made the field, are barred from postseason play this year and could not go. The ACC (Louisville) and American Athletic Conference (SMU) lose out... unless the tournament takes one extra at-large team from those conferences in their stead.

Hence Syracuse (ACC) and Tulsa (AAC) got bids they probably didn't deserve, as a restitution payoff to the power conferences for their powerful but banned programs not getting in.

Yes, this is terribly unfair from a competitive standpoint. Yes, the committee ideally should take teams on their own merit rather than out of loyalty to a conference or program. But as long as the NCAA fills a field by hand picking teams at-large, this is always going to happen with the bubble teams. Teams are always going to get seeded higher or lower than they should, or play at an out of place region when circumstances would have allowed them to play closer to home.

Unfortunately, when you give a group of rich white men the power to hand pick competitors for a championship, political and business interests become just as important a factor in your program's fate as your strength of schedule or win loss record.

Monday, March 7, 2016

If Roy Williams want automatic NCAA Tournament bids for regular season conference champs....

Roy Williams has stated for years that he feels a conference's regular season champion should get the conference's automatic tournament bid, rather than the conference's tournament champion. Should the regular season champion get an automatic tournament bid? Or the conference tournament champion?

My answer? Yes.

If you're going to do it, both should get it. We should do away with at-large tournament bids, give every conference two bids and expand the field to 72.

Much like how international soccer leagues' non-premier divisions hold playoffs to determine who gets promoted, I think the conference tournament should be played among every runner up in the conference as a winner take all event. The conference champ gets an automatic bid, and every other team in the conference plays each other for the right to the other automatic bid.

With 36 conferences, this leaves us with 72 tournament entrants. You may do one of three things.

1. Hold eight play-in games at Dayton over Monday and Tuesday to determine the 15 and 16 seeds. This gives the lowest seeds a chance to at least win a tournament game before being fed to the #1 and #2 seeds on Thursday/Friday.

2. Hold a pair of 4 team play in tournaments at Dayton on Monday and Tuesday among the lowest 8 rated teams in the field, to determine the lowest #16 seeds for Thursday/Friday's round of 64. This gives the field of 64 a little more stuffing, since more of the lowest rated teams would knock each other out during these play ins.

3. Take a page from the conference tournaments and "stepladder" the rounds with byes to reward the higher seeds. Perhaps the top 8 teams (all the #1 and #2 seeds) get a bye to the round of 32, while everyone else has to play on Thursday and Friday. This jacks with the seeding, with 18 teams now in every region rather than 16, plus instead of 1 vs 16, 2 vs 15... 3 plays 18, 4 plays 17... down to 10 playing 11... while #1 and #2 get the 1st round off.

The obvious issue with this more objective format for determining tournament teams is that you involve fewer of the power conference teams, which not only dilutes attendance and viewership, but also doesn't allow these conferences to rake in extra tournament money. That alone will likely prohibit the NCAA from ever eliminating the placement of at-large teams from the tournament.

Still, if Roy Williams is serious about giving the conference champ an automatic bid, there would need to be a carrot for the conference tournaments, which obviously won't ever go away. This would be the most workable solution, outside of that silly 96 team expansion idea the NCAA had a few years ago.

Monday, February 8, 2016

A new look at MLB realignment, six years later

Six years ago I posited how MLB would handle expansion if they added two new teams. I figured they would follow the NFL's lead and switch to eight 4-team divisions, with the playoffs consisting solely of each league's four division winners.

That would still work today, but since that 2010 post, MLB made the move to two 15-team leagues with perpetual interleague play, and Houston ended up moving to the AL. This obviously renders my San Diego to the AL recommendation obsolete, though it still allows for a new expansion team in each league.

Also, while I had recommended Charlotte as a likely expansion candidate, the city has since built BB&T Ballpark, a 10,200 seat stadium that is decidedly not suitable for a larger MLB crowd. Having committed to minor league baseball, Charlotte is now out of that running.

Omaha's TD Ameritrade Park remains a suitable MLB-friendly ballpark (even given doubts about their market). But now the 2nd expansion candidate goes back up in the air. However, groundswell has grown in recent years to bring back the once-eliminated Montreal Expos, and now it seems they would be the most likely expansion candidate should MLB expand to 32 teams.

All this throws a wrench in the realignment plan I originally proposed in 2010. So I gave it a look and realigned my realignment plan, which again focuses on geographically grouping teams as best as possible to reduce overall travel time within the division. And again, the playoffs are division champs only, making the division winner-take-all and eliminating the somewhat unfair roll-of-the-dice wildcards.

Here is how a solid realignment would look today:

AL West - Seattle, Oakland, Anaheim, Omaha

Texas and Houston would move to a more regionally accessible division, while the expansion Omaha club would join the remaining AL West teams. This unfortunately means the division has a long-travel partner, but the AL and NL North divisions were both quite loaded and no other AL teams were in easy reach of the three incumbents. So the expansion team unfortunately drew the short straw. However, travel between the West Coast teams and Omaha is a little short than travel to either Texas team, and to a less crowded airport hub, so that helps.

This division will feature three relatively competitive rivals, which for poor Omaha means their early tenure will most likely mean some 90-110 loss seasons. You also lose the hitter friendly Texas parks, swapping in Omaha's hitter friendly TD Ameritrade Park but still making this a pitcher-friendlier division.

AL North - Minnesota, Chi White Sox, Detroit, Cleveland

Most of the AL Central remains intact in the new AL North, losing only World Champion Kansas City (and boy did none of us see THAT coming in 2010). Given each team's current state of transition and development, this division would be wide open for the taking among four fairly competitive rivals. Losing hitter friendly Kauffman Stadium with KC makes this division a little pitcher friendlier, with only the White Sox's US Cellular Field being a hitter friendly park in the division.

AL South - Houston, Texas, Kansas City, Tampa Bay

Both AL-West-departing Texas teams would join World Champs Kansas City and the AL-East departing Tampa Bay Rays in a new and quite competitive AL South. Houston of course has become young and quite good, KC is now quite good, Texas has had some struggles but is still fairly good, and Tampa Bay even on a downswing is still at least a .500-caliber team. This may become the new toughest division in MLB.

On top of that, Rangers Ballpark is a launchpad, Kauffman Stadium in KC is hitter friendly, and upcoming changes to Minute Maid Park in Houston will make it more hitter friendly. If not for Tampa Bay's crappy but slightly pitcher friendly Tropicana Field, this would be the Launchpad Division. Hitters will love it, the division becomes a pitcher's nightmare, and each team's hitters could expect a sizable boost in their season stats.

AL East - Boston, NY Yankees, Baltimore, Toronto

The AL East loses only Tampa Bay (who heads to the more regionally friendly AL South), and remains one of baseball's toughest divisions. The loss of the wildcard turns both the South and the East into two of baseball's best pennant races. Plus, all four teams play in hitters parks, and Tampa's pitcher-neutral Tropicana Field goes away. Pitching just got a little tougher in the AL East.

NL West - San Francisco, LA Dodgers, San Diego, Arizona

With the past realignment, San Diego no longer needs to jump leagues, and a revised NL West only loses regionally unfriendly and largely uncompetitive Colorado. The winner take all division now gives greater weight to the always competitive Giants-Dodgers rivalry. Also, losing Coors Field makes this one of the most pitcher friendly divisions, with only Chase Field in Arizona not being a strong pitcher's park.

NL North - Milwaukee, Chi Cubs, Colorado, Pittsburgh

Cubs-Cards fans may howl at splitting up the two bitter rivals. But St Louis is more of a southern team, and fits better geographically in the new NL South. Plus, it also makes it possible that the Cubs and Cards could meet in the playoffs, whereas sharing a winner take all division means a playoff rematch would be impossible.

In any case, Pittsburgh and the Cubs' rise to power with young talent have made this a tough division, which leaves the Brewers and Rockies in a difficult position.

This becomes a very hitter friendly division, with park-neutral Milwaukee being the only non-hitter's park.

NL South - Cincinnati, St Louis, Atlanta, Miami

A new NL South would pair Cincinnati with three southern teams. In terms of travel and maintaining as many rivalries as possible, bringing Cincinnati is probably the best arrangement for MLB. A rebuilding Reds team joins the rebuilding Braves, the middling but improving young Marlins, and the strong veteran Cardinals. This division is probably all St Louis for now, as long as their veteran core remains strong. But should the Cards age and decline, this division could be wide open.

It's also one of the more pitcher friendly divisions... definitely not on the scale of either West division, but Busch Stadium plays pitcher-neutral, Turner Field (ATL) plays pitcher neutral, and Marlins Stadium (MIA) is neutral, while Great American Ballpark (CIN) is the only hitter's park. Given how many hitter friendly divisions we end up with, pitchers may prefer this division.

NL East - Philadelphia, NY Mets, Washington, Montreal

The reborn Montreal Expos would fit nicely in the revised NL East, joining the Mets, Phillies, and the former Expos now known as the Nationals. Montreal's new team would have it somewhat better than expansion Omaha, as Philly is not doing well, the Mets are good but still improving, and the Nationals while very talented have underachieved. Montreal would still enter as the likely doormat, but they'd probably be a little more competitive.

This is also the most diverse mix of parks in any division. Philly's park is a hitter's launchpad, Citi Field is a bit pitcher friendly, Nationals Park is neutral, and while it's anyone's guess how Montreal's new stadium would play, they'd likely play for now in weird and hitter friendly Stade Olympique.

------

So that's what a solid MLB realignment would look like if MLB quickly added two ready-to-play expansion teams to expand to 32. Losing the Wildcards would give added weight to each division's pennant race, and each smaller division would take on intriguing personalities of their own.

Friday, February 5, 2016

MLB's qualifying offer system needs to change, but it doesn't need to go.

MLB's qualifying offer system is the latest attempt to even the playing field for smaller market teams who can't afford to retain their free agents. However, like the Type A/B system that preceded it, the system is ham fistedly simple. You can submit a default qualifying contract offer of one season for about $15M, and if the player declines it to sign elsewhere, you get a sandwich draft pick (end of 1st round) and the signing team loses their top qualifying draft pick.

However, this vastly discourages teams with departing free agents from making the qualifying offer, since they now add a relative ton of money to the payroll if the player re-signs with them (which they may not want). It also vastly discourages other teams from signing such a player since they lose a prized draft pick if they do.

The system also has rules in place preventing offers to midseason acquisitions, which increases the organization hit to a team acquiring such a player, since they can't recoup a draft pick.

Agent-friendly baseball writers recommend the system be scrapped entirely, forgetting why such a system exists in the first place. Prior to its existence, MLB's richest teams stockpiled talent at will and left lesser teams to flounder, watch their developed young talent leave via free agency, or both.

What the qualifying offer system needs is nuance. Neither the Type A/B system or the QO had much nuance at all. Type A/B somewhat arbitrarily attached a label to a subset of free agents that could net sandwich picks, while the QO gives teams a single blunt instrument offer to levy in exchange for a draft pick. And both systems heavily penalized the drafts of any team that signed such a player.

First of all, the blanket 1st/2nd round sandwich picks and 1st/2nd round draft penalties need not be so ham fisted. It should be possible to open up lower picks in the draft, from the 3rd round down to the teens, to exchange and compensation. The expected value of these picks are far lower, and would levy a substantially smaller penalty to teams that sign a qualifying free agent... yet still aiding the draft of an organization that loses a player.

I think the $15 million qualifying offer is not a bad number... as a maximum. You should be able to offer smaller qualifying offers, that net lesser draft picks in return should the player sign elsewhere.

This can be determined by, say, average WAR per draft pick in each round over, say, the five years prior to the last six years (the span of a team's initial control over a prospect). For this season that period would be 2006-2010. This allows a complete picture of the recent relative value of picks made in that round.

Using WAR as an approximate barometer, you can make the qualifying offers relative to that $15 million total.

For example, a $15M qualifying offer would cost a team their highest available pick (top 10 picks are protected, so if a team's 1st selection is protected, they'd lose their 2nd pick). But then a team could offer, say, $7M, and if that player signs elsewhere the new team could lose their 3rd round pick. The compensated team gets a sandwich pick for that round. Or, you could offer $6M, and losing the player nets you a 4th round pick while the signing team loses theirs, and so on.

Eventually, you'll want to set a minimum qualifying offer amount, maybe $2 million or so, but you could go as low as the 10th-20th rounds in terms of compensation if you wanted, depending on how much of a qualifying offer is levied.

If the signing team has already lost their pick in the relevant round, they can exercise one of two options:

- Forfeit the next round's pick as well in this draft. If you're slated to lose a 4th round pick that you've already lost, you can choose to lose your 5th original pick as well. (Any sandwich picks you have gained are protected, so if you got a 4th round sandwich pick while losing your original 4th round pick, you won't lose the sandwich pick)

- Defer the lost pick to a future draft, losing the relevant pick in the next draft after this coming one. So in this case you could just decide to lose your 4th round pick next year.

This would soften the blow to your draft in signing a QO player. And it would make sure the team losing such players could offer more flexible QO's and get some compensation for more of their losses. A team rebuilding could load up on picks, without ruining the draft of whoever signs their departing free agents. It's much easier to lose a 3rd or 5th rounder than your top pick.

I realize this would dramatically increase the number of qualifying offers, and would send draft picks and draft orders flying all over the place. This would level the playing field, and also better encourage teams to sign these players, rather than discouraging them as the current system does.