The Hardball Times -- Alex Zelvin

Projecting innings pitched for individual games

Alex Zelvin — 2010-09-16T05:06:15+00:00

Whether you’re trying to select a pitcher for a daily salary cap format or evaluating which pitcher to use in a traditional weekly league, it would be really useful to be able to accurately forecast the number of innings a pitcher will last on any given day. Not only would that give you some idea of how much impact his performance that day will have on his rate statistics, but it also is part of the equation for determining how many strikeouts to expect and how likely he is to get the decision in the game. It’s also a topic that I’ve never seen addressed in any detail.

What makes projecting innings pitched for an individual game interesting is that expected performance will have an impact on expected innings pitched. It’s easy to mistakenly think that because most pitchers average somewhere between 5.5 and seven innings pitched per game, it wouldn’t be worth the trouble of trying to forecast innings pitched. But that’s misleading, because it averages out performance in a variety of situations. Taking all of those into account, we might be looking at a range of expected performance that goes from around four innings pitched at the low end (a bad pitcher against a good offense in a hitter's park) to maybe eight innings pitched at the high end (a good pitcher with a high pitch count limit against a weak offense in a pitcher's park). That makes the attempt to forecast innings pitched seem much more worthwhile!

In this article, I’m going to build a simple model for forecasting or projecting how many innings a pitcher will last in any given game. I’ll keep things simple, focusing on the impact of the opposing offense on innings pitched. I hope to follow it up in the future with another article that tests the effectiveness of the model. Depending on the results, maybe I’ll tackle some of the other factors that impact innings pitched too.

The majority of the time, a pitcher will come out of the game for one of two reasons. Either he's removed because he's pitching badly, or he's removed because he's reached some sort of pitch count or innings pitched limit determined by the team. There are other reasons such as injury or being removed for a pinch hitter, but these generally play a relatively minor role. When looking at a pitcher’s average innings pitched in past games, we’re looking at data that takes both major factors into account. It reflects both the effectiveness (and efficiency) of the pitcher in past games, as well as any pitch count or innings pitched limits he's been subject to.

Likewise, when looking at the average innings pitched by starting pitchers against the opposing team, we’re looking at data that reflect all aspects of how the hitters on that team have impacted opposing pitchers’ innings pitched. While it won’t be as precise as it would if we created a model based on the specific hitters in today’s lineup, we’re trying to build a simple, usable model…not a perfect simulation.

What I’m going to propose is laughably simple, but I suspect that for most situations, it’s going to be good enough to have value for many people. I’ll need three data inputs: average innings pitched for the league (NL or AL), average innings pitched of starting pitchers against the opposing offense, and average innings pitched for the starting pitcher. Basically, I’m going to create an adjustment factor by dividing the average innings pitched of starting pitchers against the opposing offense divided by the league-average innings pitched. This value (which will average one) will be used to multiply the average innings pitched of the starting pitcher being evaluated. That’s the IP forecast for today’s game.

Yes, there are all sorts of other factors that would make this a more accurate model:
-Which players are in today’s lineup?
-What are the effects of the park?
-What are the effects of being home or away?
-Should we use regression to the mean on the pitcher’s average innings pitched?
-Should we weigh the pitcher’s recent performance more heavily?

However, I believe that even the incredibly simple model I’ve described would provide value. For my next article (which should be in about a month), I’ll do some back-testing comparing the accuracy of using this model to forecast innings pitched rather than simply using the pitcher’s average innings pitched without taking the opponent into account.

Click here to learn about THT's download subscriptions.

Why daily fantasy sports contests are a better investment than the stock market

Alex Zelvin — 2009-06-03T06:18:15+00:00

I firmly believe that daily fantasy sports contests are a better investment than the stock market. Actually, I should clarify that. I firmly believe that for someone who has had an overall winning record in daily fantasy sports contests, they are a better investment than buying and holding a portfolio of stocks in the future. Obviously, fantasy contests of any sort are not a good investment for losing players. And other than in cases where sites offer "freerolls" or "overlays" to generate new business, daily fantasy contests will be a negative sum game for the "average" player, while the stock market is probably a positive sum game.

So what exactly am I saying? I’m saying that daily fantasy contests have lower variance than buying and holding a portfolio of stocks. That means that your past results give you a much better idea of whether you’re making good picks in daily contests, and that your future performance will be a lot more consistent. If you’re a winning player, you can count on a much higher percentage of winning days, months, and years than in the stock market, and the downswings should be much smaller relative to the growth of your bankroll.

To make any kind of fair comparison, we need to set up some parameters. For the stock market, I’m talking about a portfolio of U.S. common stocks. The best comparison to that in the daily fantasy world would be playing a bunch of heads up contests each day, with similar (but not identical) lineups. Each day, each stock may go up or down. The various stocks in the group will show moderate (but far from perfect) correlation with each other in their daily performance. Each day, you may win or lose each fantasy baseball contest. Your results in each contest on the same day will show moderate (but far from perfect) correlation with each other.

Let’s look at stocks first. What percentage of days will my portfolio of stocks go up? I don’t have the data available, but I suspect it’s around 50.5%. What percentage of months? I’m going to guess around 52% or 53%. Years? This one I actually remember reading about … the U.S. stock market has gone up in 57% of years. That’s an old statistic, but probably still not far off.

How about fantasy contests? What percentage of days will I come out a winner? Let’s assume that I’m a very good player, going up against average competition. I’d guess that I’m coming out ahead at least 55% of the time. If that’s the case, and I’m playing almost every day, what percentage of months will be winners? I think estimating 75% is conservative. Years? Again being conservative, I’m going to say 90%. I suspect the actually number is above 95%. Even the best stock pickers would have trouble getting that kind of results.

Assuming that I’m right about these percentages, the question is why this would be the case. Do daily fantasy contests have some characteristics that the stock market lacks that make them easier for skilled players to beat? I think they do. And I think that those characteristics have to do with what makes markets of all sorts more or less "efficient." Here are the three factors that I think going into creating an inefficient, or easily beatable market or game:

New markets: Daily fantasy contests have only been around for about two years. Most of the people who will ultimately be most successful at them probably don’t even know they exist yet. The stock market has been around for hundreds of years, and many of the best and brightest people spend their lifetime studying how to select stocks that will be winners. In other words, daily contests provide weaker competition.

Closed markets: Each daily fantasy contest is a "closed market" in the sense that entry is limited to a fixed number of participants. Once two people are entered in a heads-up contest, nobody else can enter that contest. That means that sometimes you’ll find yourself in a contest against only weak participants. In the stock market, stronger "competitors" can always get involved.

No Scalability: The size of "bet" that can be made in each fantasy contests is limited. Each participant in a $33 contest can only invest $33 in that contest. In the stock market, "bet size" is theoretically unlimited. That, combined with the openness of the markets, means that a single person with unlimited funds and omniscience can theoretically remove ALL of the inefficiency or profit opportunities.

On a separate note, I'd like to invite readers to take a look at the new site I launched this week in conjuntion with Dave Hall of Rotoguru. The site is Daily Baseball Data, and will showcase a variety of tools for players of fantasy baseball formats that use daily transactions. The initial three tools are:

1. MLB Weather Dashboard - Hour by hour forecasts for all games displayed on one screen.

2. Batter vs. Pitcher Report - Showing history of matchups for all of the day's games.

3. Sortable Statistics - For a variety of daily transaction contest formats.

Click here to learn about THT's download subscriptions.

Daily fantasy baseball contest picks for Draftbug and Snapdraft for Wednesday, May 20th

Alex Zelvin — 2009-05-20T04:30:15+00:00

Like last week, I'm doing an in-depth analysis of Wednesday's games for players in daily fantasy baseball contests. At each position, I’ll list the top five healthy players according to my ratings, along with their prices in Draftbug. Snapdraft’s scoring is VERY similar, so the same ratings should apply, although the pricing varies enough that you could end up with a very different lineup using the same ratings. Although the discussion will focus on Draftbug, the same thought process applies in Snapdraft, and I'll include a lineup for Snapdraft at the end of the article as well.

For each position, I’ll list my ratings for the top five players, as well as any others that I seriously considered putting in my lineup for the day. Depending on the size of the contest and the buy-in, the salary cap varies between contests at Draftbug (unlike Snapdraft, which always uses the same salary cap). One change from last week is that I'll use a salary cap of 1400 points (instead of 1500). Not only does that require more interesting tradeoffs and compromises, but it is the cap used for the daily freeroll contest, which is likely to be most people's first exposure to the contests. Keep in mind that the ratings listed are specifically for Wednesday’s games, based on factors such as park, opponents, platoon advantage, and home field advantage.

One reader last week finished second in several contests using my ratings, and concluded that "the system works more or less as advertised." While it's certainly gratifying to see someone do well using information I've provided, it's probably worth mentioning that it takes a while to really know how good you are at these contests. In the long run, they're very dependent on skill ... more than most fantasy baseball formats. However, in the short run, there's extremely high variance in the results.

Starting Pitchers:

1 Vazquez     15.4     199
2 Kazmir      14.9     249
3 E. Santana  14.8     249
4 Gaudin      14.2     N/A
5 Gallardo    13.9     209

This is an incredibly easy choice. I typically am willing to spend whatever it takes for the top starting pitcher, but in this case Vazquez is relatively cheap, and he's actually probably underrated by my system's incomplete ability to project innings pitched.

Relief pitchers:

1 Papelbon  3.9   210
2 Fuentes   3.9   156
3 Rodney    3.5   102
4 Gonzalez  3.5   96
5 Qualls    3.5   90

In general, in a large multiplayer contest (like the freeroll), I prefer taking a starting pitcher and and reliever from the same team. However, Vazquez tends to pitch a lot of complete games, so I'm going to use Qualls instead of Gonzalez.

Catchers

1 Martin    .80  174
2 McCann    .69  186
3 Mauer     .67  174
4 Laird     .61  72
5 Napoli    .60  120

For hitters, my ratings reflect their per-at-bat projection. That means that I'm penalizing players who tend to get more at-bats—those who bat early in the lineup, play on teams with high on-base percentages, and don't get removed late in games. As the season progresses, I'll build some kind of estimate of plate appearances into my statistical model, but for now it's something that I just need to keep in mind when I'm making my picks. The only two reasonable choices at catcher today are Martin or Laird. I'll be using Martin. With the restriction of the 1400 salary cap, he and Vazquez are the only two top-rated players that I'll be using.

First Base

1 Cabrera    .94  288
2 Pujols     .82  324
3 Teixeira   .78  176
4 Morneau    .72  240
5 Loney      .71  144

Cabrera has such a strong projection that I had hoped I'd be able to use him today. Unfortunately, I wasn't able to find the room under my salary cap, so I'll be taking Loney to save salary. When I'm evaluating potential lineups, and I seem to be falling just barely short of being able to field a substantially better team, I often look for additional contests to enter that use a higher salary cap. Likewise, if I have a lot of room left under the salary cap (which doesn't happen as often), I look for contests with a lower cap.

Second Base

1 Pedroia   .90  234
2 Polanco   .73  102
3 Utley     .72  252
4 Hudson    .71  96
5 Cano      .70  174

Like Cabrera at first, I had hoped I'd be able to afford Pedroia. I couldn't, so I'm taking Polanco.

Third Base

1 Rodriguez  .86  300
2 Wright     .86  306
3 Jones      .83  210
4 Lowell     .76  126
5 Blake      .69  108

Lowell is a great bargain at third base today. Note that the gap between his projection and the top third basemen is much less than that between some of the other "bargain" players and the top players at their positions. It's becoming clear that my model "likes" the Yankees, Dodgers, Red Dox, and Tigers hitters today. But keep in mind that not all players on a team will be affected equally by a match-up. Factors such as platoon advantage, stolen base potential against pitchers who don't hold runners on well, ability to draw walks against pitchers with poor control, and more can result in some surprising rankings among players on the same team.

Shortstop

1 Ramirez     .83  324
2 Reyes       .83  318
3 Furcal      .81  210
4 Jeter       .65  174
5 Rollins     .65  288
  Aviles      .64  138

In Furcal, I was able to afford I player not far from the top two, for a substantially lower price.

Outfield

1 Beltran    .83  234
2 Ordonez    .82  174
3 Granderson .81  192
4 Bay        .81  210
5 Kemp       .80  210
  Drew       .80  96
  Pierre     .79  84
  Ethier     .76  154
  Damon      .74  138

I tend to look for bargains among outfielders. In general, there are more good alternatives among outfielders most days. Today is no exception, as Drew, Pierre, and Ethier (or Damon) provide good, cheap alternatives who are just barely worse than the top five outfielders. Several of these players (Pierre in particular) seem to be ending up on my team and my opponents' teams often enough that they're likely to be among those whose prices will be modified soon.

Using the same ratings, my Snapdraft team for the day is Martin, Cabrera, Pedroia, Rodriguez, Furcal, Beltran, Drew, Pierre, Vazquez, and Kazmir. That looks a lot more like an all-star team, and that's fairly typical. Snapdraft uses the same $27M salary cap for all contests. The disadvantage is that teams in the same contest tend to look fairly similar, because it doesn't force tough decisions. The advantage is that you can use the same lineup in multiple contests at the click of a button.

Click here to learn about THT's download subscriptions.

Daily fantasy baseball contest picks for Wednesday, May 13th

Alex Zelvin — 2009-05-13T05:39:15+00:00

I’m going to try something a little different this week, and discuss my daily contest strategy for Wednesday’s games in detail. I think that should give those who haven’t tried the games yet a better feel for what’s involved in them, and prove useful for those who are trying to plan their Wednesday contest lineups. If feedback is positive, this is something I could do on a regular basis.

At each position, I’ll list the top five healthy players according to my ratings, along with their prices in Draftbug. Snapdraft’s scoring is VERY similar, so the same ratings should apply, although the pricing varies enough that you could end up with a very different lineup using the same ratings. In any case, the thought process for making selections would be the same regardless of what daily contest format you're playing.

A few people have asked whether I’d be willing to share the "guts" of my ratings spreadsheet here. This is probably about as close as I’m willing to get right now. However, if people email me at zelvin30@hotmail.com I’m happy to discuss specific issues in statistical modeling for daily contests.

For each position, I’ll list my ratings for the top five players, as well as any others that I seriously considered putting in my lineup for the day. Depending on the size of the contest and the buy-in, the salary cap varies between contests at Draftbug (unlike Snapdraft, which always uses the same salary cap), so I’m going to use a cap of 1500 points, which is more or less in the middle of the range. Keep in mind that the ratings listed are specifically for Wednesday’s games, based on factors such as park, opponents, platoon advantage, and home field advantage.

Starting Pitchers:

1 Lee         13.8     279
2 Nolasco     13.1     209
3 Pettitte    12.3     100
4 Lilly       12.2     189
5 Outman      10.8     10

For pitchers, the points projection listed is for the game as a whole. In general, starting pitchers have the widest spread in projected points, so I often spend the salary on my top-rated pitcher. I can find other places to save salary that won’t have as much impact. This is an unusually weak day for starting pitchers, and Cliff Lee (at home) against Buehrle and the White Sox is my top-rated pitcher. He may actually be even better than my rating suggests, because good control pitchers aren’t currently given the favorable innings pitched projections they deserve in my statistical model. Despite the high price, and the unlucky start he had to his season, he’s a pretty easy choice for me.

Relief pitchers:

1 Rivera    3.9   186
2 Wood      3.6   126
3 Papelbon  3.5   210
4 Broxton   3.5   138
5 Hanrahan  3.3   78

I’m going to go with Wood for my relief pitcher. An argument could also be made for Rivera (who has the highest rating) or Hanrahan (who is substantially cheaper). One thing to note here is that Broxton’s price will probably need revising (upwards) the next time prices are adjusted, which happens every few weeks. At his current price, he’s a pretty big bargain, since he’s often my top-rated reliever. He’s one of the few players who tends to get picked by a high percentage of players at his current price level.

Catchers

1 Mauer     .78  174
2 Soto      .71  162
3 McCann    .68  186
4 Iannetta  .67  120
5 Martin    .67  174

For hitters, my ratings are listed per at-bat. Ideally, my model would incorporate average batting order position, and the team's on-base percentage. For now, I'm simply assuming about five plate appearances per game for all players, and keeping in mind that players who bat higher in the order are likely to outperform those who don't. The two reasonable choices at catcher today look like Mauer and Iannetta. I’ll go with Mauer, but if I was in a contest with a cap lower than 1500, I’d probably save some of my salary room by using Iannetta. At catcher, there’s always a risk that your players will take a day off, so the real key is figuring out who isn’t going to skip the game and score no points at all.

First Base

1 Pujols     .93  324
2 Morneau    .81  240
3 Hoffpauir  .78  28
4 Helton     .72  78
5 Pena       .72  174

Today is a fairly typical day for first basemen, with Pujols rated the highest by a lot. However, his high price makes him an impractical choice on many days. Hoffpauir is treated very favorably by the ZIPS projections that I’m basing my daily calculations on. However, he’s the type of player (young and without much major league experience) where the various ratings systems tend to disagree, and where ZIPS may not be as good as some of the others. More importantly, he’s not usually in the starting lineup, so he’s not a good option unless you can check the lineup before your picks "lock" for the day. Helton, on the other hand, is often a great cheap value, as his strengths are a good fit for the points scoring system, and he benefits from playing in Colorado half of the time.

Second Base

1 Kinsler   .81  246
2 Pedroia   .78  234
3 Utley     .77  252
4 Roberts   .72  210
5 Matsui    .71  96

Many days, one of the top three second basemen has such an edge that it’s not worth looking for cheaper alternatives. That’s because they’re not only good, but each plays in an extremely favorable home park. However, today Kaz Matsui is a good, cheap alternative. He benefits from playing at Coors field against a bad pitcher (Jason Marquis). Marquis is a particularly good opponent for players with some speed, because he’s very easy to steal bases against.

Third Base

1 Wright     .81  306
2 Jones      .80  210
3 Figgins    .79  156
4 Atkins     .76  168
5 Rodriguez  .75  300

With a number of relatively evenly matched choices at third base today, Figgins looks like the best value. One caveat here is that I’m still using the park factors for Shea Stadium and the old Yankee Stadium. I know that the early results seem to indicate that Yankee Stadium is now a much better hitters’ park, but I’m going to wait until there’s a little larger sample size before I make the changes in my model. So I may be cheating Wright, and especially Rodriguez.

Shortstop

1 Reyes       .82  318
2 Ramirez     .74  324
3 Theriot     .73  126
4 Rollins     .72  288
5 Tejada      .71  144
  M. Izturis  .71  16

Theriot is a great value at shortstop today. He’s at home in a hitters park, which helps his rating. But the real kicker is that he’s up against Chris Young, who has been one of the very worst pitchers at preventing steals for years. This is one of the cases where using analysis of component statistics really helps. Instead of rating Young as a "good pitcher" across the board, the ratings reflect the fact that he’s a "good pitcher" against many hitters, but an "awful" one against good base stealers. I listed Izturis as an interesting bargain, but like Hoffpauir at first base, you should only use him if you'll be able to confirm that he’s in the lineup for the day.

Outfield

1 Soriano    .88  234
2 Lee        .84  228
3 Crawford   .80  222
4 Cruz       .80  155
5 Beltran    .79  234
  Bradley    .76  126
  Hunter     .75  150
  Abreu      .73  150
  Spilborghs .71  36

With three outfield slots in the standard roster configuration, I find choosing outfielders to be more of an art than a science. Spilborghs is probably due for a price increase, as his low price reflects the lack of playing time he received last year. Cruz was one of the players adjusted upward in the last round of price changes, but he remains a good bargain when he’s home against weak pitchers. Bradley is also worth considering, as long as he’s not nursing a new injury or facing a suspension for anything.

I'd like to hear everyone's thoughts either via email or in the comments section, both about the specific picks for today's games, and about whether this type of article is useful or interesting.

Click here to learn about THT's download subscriptions.

Keys to winning daily fantasy baseball contests

Alex Zelvin — 2009-05-06T05:23:15+00:00

What follows are some of the keys to winning daily fantasy baseball contests, such as those offered at Draftbug, Snapdraft, and Fantasysportslive.

Look for bargains — This is kind of obvious, but the salary cap limitations in these contests mean that you’re not usually going to be able to use an all-star type lineup. You’re going to need to search out some undervalued players and use them. One common source of these is players who will have an expanded role today due to an injury to another player. However, there are players who provide good value for more extended periods of time. Sometimes this is because player prices aren’t updated based upon superior performance. Other times, these values are available because the player’s value in a particular scoring format surpasses their reputation, so nobody notices that they’re underpriced.

Check the news — Check the news to make sure that you’re not using anybody in your lineup who won’t be playing today. Lineups in these contests typically lock in after starting lineups for the early games are announced, so ideally you should make sure that your players are actually in the lineup. This also means that if you are going to be able to check lineups prior to your roster locking, then players with early games are going to be worth a few percentage points more to you on average than those with later games.

Understand the rules — This goes for any format of fantasy baseball—you should always know the rules thoroughly. In particularly, the scoring systems vary in different daily contest formats, and you should make sure you’re selecting players who fare well in the scoring system your contest uses. A player like Jason Giambi is a pretty valuable if your system uses walks, but not so much otherwise. Also make sure that you know how the rules treat special cases like doubleheaders and rainouts.

Larger contests require more risks — In a two-person contest, you should just go with the best possible lineup. In a larger contest (say 100 people), where the prize payouts are typically very top-heavy, you’re going to need to take some chances to have a shot at the top spot. Play for first place, not 10th. In the larger contests, you want to increase variance in your scoring. One way to do this is to pick players whose performance is likely to show a high correlation. For example, pick the starting pitcher and closer on one team. Another way is to increase variance is to take players whose scores tend to be more "feast or famine," such as home run hitters who strike out a lot. You also should be a little more willing to pick a clearly superior player who has a small to moderate chance of being rained out, if you’re playing in a larger contest. But in a heads-up contest, just play it safe and make sure that all your players will be playing today.

Take opponents into account — When you evaluate players for your team, consider who their opponent is today. For pitchers, how strong is the opposing lineup? How strong is the opposing starting pitcher? For hitters, how strong are the opposing starting pitcher and bullpen? I’ve built a fairly complex statistical model to do this, and I suspect that other top players have as well.

Take park factors into account — Where is the game being played? Ideally you want pitchers to be in favorable pitchers’ parks and hitters in favorable hitters’ parks, although there are plenty of cases when other factors may override this. This too, is part of my statistical model. However, like evaluating opponent quality, you can do it somewhat effectively simply by eyeballing the schedule and pitching match-ups each day before you make your picks.

Click here to learn about THT's download subscriptions.

An introduction to daily fantasy baseball contests

Alex Zelvin — 2009-04-29T05:44:15+00:00

When I started writing for The Hardball Times, the plan was for me to write about Rotohog, since that was the game where I had experienced the most success and to which was devoting the most time. That plan went out the window when Rotohog eliminated most of the prizes for its contests. Fantasy baseball is always fun, but it’s a lot more fun when there’s some money on the line, and without substantial prizes, I wasn’t as interested in Rotohog. Since then I’ve been writing on topics related to a wide variety formats using daily transactions. However, I’ve been looking to introduce a more consistent focus to my articles. It’s been decided that going forward I’ll be writing about the fantasy baseball topic nearest and dearest to my heart—daily contests.

For those of you who aren’t familiar with them, here’s how daily fantasy baseball contests work. You pick a team today. You win (or lose) tonight. That’s it, more or less. Most sites run the contests as "salary cap" games, where each player has an assigned cost, and you have to pick a full lineup without exceeding the salary cap. Scoring is based on a points system. In addition, some sites offer "live draft" format contests, where you actually do a quick draft with one or more opponents. Games are run with anywhere from 2 up to 100 (or more) contestants, and for stakes ranging from play money to $200 or more per contest.

Close to a dozen sites offer these contests, including Draftbug (which is my site), Snapdraft (the most heavily marketed site so far), and Fantasysportslive (the earliest site to offer these games). While they share many features, there are some differences in the sites' look and feel, contests, scoring, roster configurations, and other features. However, they have enough in common that the strategic issues players face will be similar on all of them, and most of what I talk about should be relevant no matter which you play on.

So why do I think these games are so great? There are two reasons.

Daily fantasy baseball contests allow a lot more room for use of Sabermetrics to gain an edge over your opponents. In a traditional league, the most important skills include performance projection for the season, player valuation for your format, and ongoing game strategy. For the first two, commercially available forecasts and ratings have made it very hard to gain any kind of substantial edge over opponents. Game strategy can provide an edge but isn’t especially dependent on understanding or using sabermetrics. By contrast, success at daily fantasy contests is almost entirely about who has the best “sabermetric-fu.” While the results of any one contest can appear to be almost entirely luck, the cumulative results of multiple contests per day over the course of the season have an extremely high degree of skill. And that skill encompasses forecasting player performance each day based on a host of factors including skill, park factors, home field advantage, opposing starting pitcher, opposing bullpen, health, weather, opposing lineup and more. Each of those offers the opportunity for a wide range of approaches, which will impact your success or failure. For numbers geeks, these games are like a sabermetric playground, where we can profit based on the success of our ideas.

On the other hand, as obsessed as I am (and many of you are) with baseball in general, and fantasy baseball in particular, one of the drawbacks of traditional (full season) leagues is that they’re a grind. Particularly the daily transaction leagues can feel like a chore after a while. If you take a day off it can really hurt your team, and a summer vacation can ruin your entire fantasy baseball season. Daily contests solve that problem. You play them when you have the time, and don’t play when you don’t have the time. For those of us with families and other responsibilities, that’s a real blessing. I may spend hours each day thinking about and working on fantasy baseball, but every once in a while there’s a day when it’s really hard to find even a few minutes for it. Daily contests eliminate the burden of having to check my lineup on those days.

I hope in the coming weeks and months I can not only show you how terrific these games are, but provide you with many of the ideas and tools that will help you succeed in them!

Click here to learn about THT's download subscriptions.

Are you smarter than a sabermetric spreadsheet? (Part 2)

Alex Zelvin — 2009-04-15T06:35:15+00:00

Last week I provided a points scoring system and asked readers to list the five outfielders who they expected to score the most points on Friday, April 10. Of the readers who replied, the top five picks were Grady Sizemore (six picks), Alfonso Soriano (six picks), Curtis Granderson (six picks), Matt Holliday (four picks), and Manny Ramirez (four picks). The spreadsheet that I use to evaluate hitters based on their matchups, park, and other contextual factors came up with a very different list, only agreeing with the readers on Sizemore. The other outfielders it identified in the top five were Carlos Quentin (two reader picks), Jermaine Dye (one pick), Shane Victorino (one pick), and Vladimir Guerrero (zero picks).

So what accounted for the difference?

It looks like the players on the readers' list tended to be really good and other than Ramirez are very well-rounded players. The computer put more of a focus on match-ups and contextual factors. It gave Quentin and Dye a lot of credit for playing at home, in a hitters' park, against a really bad opposing starting pitcher (Dickey). Victorino, too, benefited from playing in an excellent hitters' park against a bad pitcher (Marquis). Marquis is also very easy to steal bases against, which is a substantial advantage for a fast player like Victorino. I assume that was overlooked by most (if not all) readers. I certainly wasn’t aware of it until I looked over the results of the spreadsheet calculations. In Guerrero’s case, I suspect that the readers may have actually done a better job in their evaluation than the spreadsheet. The calculations treated Wakefield as a really awful pitcher based on his K/9, BB/9 and GB% rates. What they don't know is that he’s a knuckleball pitcher and that knuckleballers tend to do better than their component statistics would suggest.

Among the other popular readers picks, I suspect that Granderson and Soriano may be worse players in this scoring system than many readers gave them credit for, due to their lack of walks. That’s actually one of the things that's really surprised me since I started maintaining the spreadsheet and using it for multiple game formats. There's a drastic difference between the values of players in different games, and in some cases a player may rank very high in one game and far lower in another. Just knowing that a player is "good" isn't enough. You need to know what they're worth in the scoring system that your league or contest uses.

Comparing the readers' picks with the spreadsheet picks really points out the advantages and disadvantages of relying on calculations rather than intuition for your picks in daily contests. While the spreadsheet can often do a better job measuring and balancing a number of different factors related to the players and the context of the game, it can only take into account factors that have been programmed into the statistical model. If you haven’t gotten around to including them, it won’t know about things like knuckleballs, rain, injuries, defensive replacements, and other factors that may occur infrequently or be hard to quantify. On balance, I think a good statistical model can outperform the intuition of any expert, but the model will do best when monitored by a knowledgeable person.

While we all know that one game is far too small a sample size to really answer the question of who was "right," it’s still fun to take a look at how the players in question performed. Here are the point totals for each of the nine players listed above:

Sizemore: 1
Soriano: -2
Granderson: 6
Holliday: 4
Ramirez: 4

Sizemore: 1
Quentin: 8
Dye: -1
Victorino: 2
Guerrero: 4

Almost a tie, as the spreadsheet outperformed the people 14 to 13. A side benefit of the spreadsheet’s focus on matchups is that it will often identify players who are available more cheaply than the stars that most people prefer.

Click here to learn about THT's download subscriptions.

Are you smarter than a Sabermetric spreadsheet?

Alex Zelvin — 2009-04-08T06:41:15+00:00

I wrote a short post on my blog a few days ago about the drawbacks of relying on intuition or instinct when making picks in daily fantasy baseball leagues. I think it’s a topic that’s worthy of some elaboration. After all, if you’re knowledgeable about baseball and have a solid understanding of statistics in general and Sabermetrics in particular, shouldn’t you be able to come up with roughly the same result by looking things over carefully as you would if you actually did the calculations to take all the relevant factors into account? The answer is a resounding "no."

Using a spreadsheet or computer program is going to provide you with at least three big advantages.

The first is that once a player or a factor is incorporated into your statistical model, you’ll never overlook it. You won’t forget about the existence of a player or forget to check whether one of your hitters is facing a lefty or righty starter.

The second is that you’ll be able to work faster. Assuming all my supporting data is up to date, I can turn a schedule from MLB.com into ratings for the day and pick a lineup for almost any contest in 5-10 minutes. If I had to research the park, opponent, and other factors for each potential member of my lineup, it would likely take an hour or more.

But by far, the most important factor is that the human mind (at least most of them) simply isn’t well-equipped to evaluate a large number of contextual factor simultaneously, and is particularly badly suited to assigning the proper weight to various factors based on subtle differences of degree. As an example, are you really able to judge which is worth more in your format: a .300 hitter with 30 HR, and 5 SB, playing in a park that inflates runs and home runs by 5 percent against Johan Santana, or a .280 hitter with 35 HR and 0 SB, playing in a park that deflates runs and home runs by 5 percent against Brett Myers? Then consider all the factors I didn’t even mention—platoon advantages, bullpens, home-field advantage, and many others. It seems unlikely that anyone is capable of assessing the importance of each of them, and arriving at a reasonably accurate evaluation or projection of each player’s value for the day.

As a kind of thought experiment or game, I’m asking of each of you who reads this to take a look at the games scheduled for Friday and rank the top five outfielders. We’ll use the following points scoring system:

Single: 1 point
Double: 2 points
Triple: 3 points
Home Run: 4 points
Run: 1 point
RBI: 1 point
Walk: 1 point
Stolen Base: 2 points
Strikeout: -1 points

Post your picks in the comments section here. My own picks just based on intuition are Beltran, Rios, Sizemore, Hamilton, and Guerrero. I’m primarily looking at who the opposing pitcher is and what park the game is being played in. Sometime before Friday, I’ll put the results of my spreadsheet calculations in the comments section, and next week I’ll discuss some of the factors that influenced the calculations. It should be interesting to see how many of our picks differ from what my formulas indicate, and what factors cause the differences. While the results of one day’s games are too small a sample size to know which set of predictions are "better," hopefully we’ll still be able to learn something interesting from the exercise.

Click here to learn about THT's download subscriptions.

Favorable Opening Day schedules and matchups

Alex Zelvin — 2009-04-03T06:08:15+00:00

One of the keys to winning at fantasy baseball formats with daily transactions is to look at match-ups on a day to day basis. That holds true whether you’re in a daily contest format that allows you to turn over your entire team each day, or a more traditional daily transactions format, where you’ll be focusing on rotating some of your more marginal players between your lineup, your bench, and the waiver or free agent pool.

When evaluating the beginning of the season, matchups can be a little tricky to evaluate. While we know for sure which opposing team everyone will be facing, and which park they’ll be playing in, it can be difficult to gather complete information on starting pitching rotations. That makes it hard to fully assess how favorable the schedule will be to various hitters. I won’t even be trying to calculate daily ratings until Saturday. That said, based on what we do know, we can identify some teams and players that may be in particularly good situations on Opening Day and for the rest of the initial series of the 2009 season.

Cleveland Indians – Playing at Texas. Indians hitters will be in a good park for hitters, facing a weak pitching staff with no ace and a questionable bullpen. The only slight minus is that they’re on the road. Grady Sizemore is almost certainly the most valuable hitter for Opening Day if you play in a daily format that allows complete lineup turnover each day, and likely will be the most valuable outfielder for the entire opening series. Victor Martinez will be a clear standout at catcher as well. Players like Choo and Peralta may offer good value in salary cap leagues.

New York Yankees – Playing at Baltimore. Another situation where a road team will be facing a much weaker than average Opening Day starter. That’s going to give Yankees hitters a nice start to the season, and will also give Sabathia an excellent chance of winning his first start. Depending on the format, Teixeira may be worth just about as much as Pujols for Opening Day, at a far lower cost. Other hitters like Cano and Jeter will move way up relative to the elite players at their position, again offering potential bargains in salary cap leagues.

Arizona Diamondbacks – At home against Colorado. This is a great situation for Arizona players. At home against a team with weak hitting and weak pitching. It’s made even better for their hitters by the fact that they play in a good hitters park. Like Sabathia, Webb should have an excellent chance of picking up an Opening Day win, and also like Sabathia, Webb is efficient enough that he may pitch deeper into the game than most starters are able to at the beginning of the season. I don’t consider any of Arizona’s hitters really elite, but in salary cap games there may be one or two who offer good value given the situation they’ll be facing in their opening series, and in Yahoo leagues it’s worth taking a look at any unowned hitters on their team who might be useful for a spot start in your lineup.

Speaking of Yahoo leagues, several weeks ago I discussed the value of relief pitchers with starting pitcher eligibility. At this point, it looks like there are at least three who could have substantial value at the beginning of the season. Hong-chih Kuo I’ve already discussed. He’ll contribute wins, strikeouts, and excellent ratios, while filling a starting pitching slot on your roster that will rarely be used for an active player otherwise. With the Seattle closer job still open, and Brandon Morrow not yet able to pitch more than a few innings, I think there’s a good chance that Morrow will move back into the closer’s role for the Mariners. And the Brewers have already announced that with Trevor Hoffman out to start the season, Carlos Villanueva will get first shot at being the fill-in closer.

Click here to learn about THT's download subscriptions.

Using high variance poker tournament strategies to win money at fantasy baseball

Alex Zelvin — 2009-03-26T07:22:15+00:00

If you’ve spent any time talking to successful poker tournament players, most of them will tell you that in large multi-table tournaments with hundreds (or thousands) of players, high variance strategies are one of the keys to success. That doesn’t mean taking stupid chances, but it does mean that in a large field contest with top heavy prize distribution, if you have two options with similar expected value, then the higher variance option is usually going to win you more money in the long run.

The same concept can be used in fantasy baseball contests that share the characteristics of being large (in terms of number of contestants) and have top heavy payouts.

The ideal format to apply this idea is the daily fantasy contest sites, such as Draftbug (my site), Snapdraft, and Fantasysportslive. With contests lasting just one day, high variance strategies are easy to apply, and probably appropriate for any contests with 10 or more participants. While sacrificing expectation for variance doesn’t make sense for the very best players in longer duration fantasy sports contests, it may be beneficial for even the best players in contests lasting just one day, since even the very best players will only come out ahead very slightly more than their "fair share" of the time.

However, there are situations where high variance strategies can be applied to longer duration "global contests" such as ESPN, Rotohog, and Sporting News Salary Cap Challenge and Ultimate. The key is that these are all large contests, with thousands of entrants. The odds that you’ll win simply by making better picks are very small.

The basic idea behind sacrificing "expected points" for variance is that prize distribution is so skewed towards the first place finisher in most of these contests, that if you can increase your chances of coming in first more often, it’s worth reducing your chances of finishing in other positions in the top half of the field. A 10 person daily contest doesn’t reward you for finishing fourth instead of last, and a 10,000 person contest probably pays little or nothing for 300th place. If you want to win money in the long run in these contests, you need to shoot for the top.

One way to accomplish this is to choose players whose performance is likely to be strongly correlated. The easiest way to do this is in daily fantasy baseball contests is to choose players on the same team. If they knock the opposing starter out early and get to face the dregs of the opposing bullpen, that's going to benefit both of them ... leading to a positive correlation in their scores for the day. If the batter hitting fourth gets an RBI, there's a pretty good chance that the batter hitting ahead of him got a run ... again leading to greater correlation among their daily scores. Most daily games won't allow you to choose all players from the same team (because that would legally count as sports betting), but choosing mostly players from the same team is definitely a good idea in larger contests ... particularly if you can identify a bad opposing starting pitcher to go against. For games that give pitchers large amounts of points for wins and saves, you can greatly increase variance by selecting a starting pitcher and closer from the same team.

While this isn’t as effective in longer duration contests, there are some special situations that will allow you to get high "internal correlation" among players over the course of the season. For example, if you believe (as I do) that the Mets’ new home park (Citi Field) is going to be an extreme pitchers park, you can select as many Mets pitchers as possible. Either you’re wrong or you’re right … but if you’re right, it’s going to help all of your pitchers.

The second way to increase variance in fantasy sports contests is to differentiate from your opponents. This is a little trickier. In order for it to be a viable strategy, you need a few conditions to exist. Score should be heavily influenced by one player ... for example a game format where a single starting pitcher generally scores almost half of a team's points for the day. You need to know that most of your opponents are likely to choose the same player for that position. Imagine a game with 20 contestants where the entire score is derived from a single starting pitcher. Now imagine that there are only two starting pitchers available today ... Jake Peavy and Mike Pelfrey. Who is the better pick? Almost certainly Pelfrey! While Peavy might have a 75 percent chance of winning, if you win you'll be sharing your first place prize with about 18 other people. If Pelfrey wins (25 percent chance) you're likely to win the entire prize. This is a great (although admittedly extreme) example of how differentiation can help you in these contests.

This too is an easier strategy to implement in daily contests. Typically the only place for differentiation in full season contests is if you’re trying to make up ground in the last few days or weeks of the season. However, you can use the same "special situations" like the Citi Field example above to differentiate your team from others. Another example would be in a contest where you need to balance roster value with playing match-ups effectively. If you know that everybody else is going to focus on building roster value, maybe you can win it all by focusing on match-ups at the expense of roster value. The odds are that the masses are correct in their approach. But if you go along with them, your chances of winning are still miniscule. If you take a different approach, and that approach works, you may be the only one in position to win it all.

Click here to learn about THT's download subscriptions.