1. To begin, why don't you tell us about WhatIfSports and what it is that you do.

WhatIfSports.com began in 2000 as a way to settle the age-old, bar-room sports debates. We are located in Cincinnati, a hotbed for college basketball. We often argued about which of the great Cincinnati, Ohio State, Kentucky, Louisville, Indiana, Dayton and Xavier teams were the best until we decided to find an answer.

Today, the site allows users to play hypothetical matchups of professional baseball, basketball, football and hockey and collegiate football and basketball teams from throughout the history of each sport by using the free SimMatchup simulations. Want to know who would win between the 1998 Yankees and the 1927 Yankees? We can tell you (and we will tell you in our next feature).

Registered users can also build “Dream Teams” of their favorite players of all time in those sports for free. Want to know how an outfield of Babe Ruth, Mickey Mantle and Joe DiMaggio would look? You can do that here (we’ll let you decide who plays center). Those Dream Teams can also be submitted into SimLeagues where the owners of those teams have the opportunity to play GM and manager of the team for full seasons, while competing against others for pride and prizes.

And finally, we have added games that involve fictional players where users get to manage every aspect of their own baseball franchise, college basketball or football program or stockcar team (more on this later).

My title is Quantitative Analyst and Content Manager. Essentially, I am in charge of managing the research and statistical analysis that goes into our simulation engines as well as creating and writing the content that comes from it. It’s pretty fun.
 

2. How does WhatIfSports differentiate itself from other baseball simulators like Diamond Mind and Strat-o-Matic?

We listen. This is not to say that our competition does not understand its user base, but I feel that our greatest asset is that we are always listening to our customers and willing to do whatever it takes to improve our games based on what they say or want. For this reason, our website is just as much of a community of sports fans – with active forums and intense competition - as it is a gaming or content site. Also, because of our willingness to listen, our customer service is second-to-none. The developers of the games handle many of the tickets, and every ticket is reviewed within 24 hours.

Obviously, having eight unique products, spanning five different sports – not to mention the free SimMatchup and Dream Teams – sets us apart from any other site. Users of our site will find a customer-friendly, consistently-updated site with almost limitless options, where they can be as involved as they would like and always have a say.

It’s tough to beat our prices across all of our products as well.
 

3. Can you give us some examples of some real-life simulations (baseball or non-baseball) in which WhatIfSports accurately predicted the outcome of a game, match, or season?

No problem. We have actually been on a bit of a roll lately. The content part of our site (i.e. projecting forward and telling people about it) has only been active since late 2005. Our best predictions by sport since then include: 2006 WS – Cardinals in Five, 2008 NCAA Basketball Tournament – Kansas over Memphis (as of the day after Selection Sunday), 2008 NBA Finals – Celtics in Six, 2008 Stanley Cup – Red Wings are overwhelming favorite (51% of 16 teams in Playoffs), 2006-07 College Football – 88% bowl winners and 2007 Super Bowl – Colts over Bears 26-20.
 

4. With baseball specifically, what types of stats are used in your simulations and which carry the most weight (if you can tell us that)?  Do you use sabermetric-friendly statistics?  How do the simulations actually work?

I could probably write/talk for several hours about any of our engines, but I will try to make this as succinct as possible. Baseball is actually the easiest simulation engine to write because everything is event based as opposed to continuous flow like basketball and hockey. Just about everyone who plays a role in one of those events in baseball can be quantified easily, unlike football – let alone the fact that stats have been kept (and kept well) in baseball for over 100 years.

In baseball, the main event is the plate appearance. To simulate a plate appearance, we break it down into steps where if one result is achieved we advance to a different step than if another result is found – like a “choose your own adventure.” Assuming a normal game situation with no one on base, we look at the batter’s walks per PA and the pitcher’s walks per batter faced to determine the likelihood of a walk. A random number is drawn. If it is within the constraints of the probability, a walk occurs. We do the same with hit by pitch.

If neither occurs, we move on to whether there is a hit or an out. Ballpark factors, lefty vs. righty splits and specific weighting from extreme hitters and pitchers comes in to play here, but essentially this is the hitter’s batting average compared to the pitcher’s average allowed. If a hit occurs, we determine the type of hit based on the weightings of the hitter and pitcher’s types of hits per hit (for or allowed). Historical ballpark factors act as modifiers to the likelihood of the type of hit. Once we know the type of hit, we can determine where the ball was hit to see if the defender at that position is capable of turning a hit into an out with superior range. If an out is selected, we first determine whether it was a strikeout. If it is not a strikeout, an out can turn into a hit by a player with inferior range. It can also lead to a player getting on base due to error.

The process includes our interpretation of “log5 normalization,” a method popularized by Bill James. I will discuss the normalization aspect in the next question, but because weights are referenced in this question, I want to talk about a fundamental flaw in an assumption to James’ approach that I believe we have corrected. One of the principle assumptions to log5 normalization is that the batter and pitcher have equal weight in the plate appearance event (at least before the defense and ballpark affect it). Our research indicates that this is not true for any step in the event tree. Each of the steps has a different pitcher vs. hitter ratio, but in EVERY case, the hitter has more significant weight on the outcome than the pitcher. This includes the probability of a walk, which favors the hitter, yet is close to an even split. The best reason for this phenomenon that we can surmise is that the batter always has ultimate control over whether or not he swings. It does not matter how wild a pitcher is if the batter swings. Vladimir Guerrero can help us understand this.
 

5. Can you run historic simulations, maybe putting a team from 50 years ago against a current team?  How do you equate statistics across time periods?

Yes, we include the ability to simulate games involving any team or player from 1885 to 2007. To handle this, we apply normalization to the process above. A player’s inputs to the plate appearance event can be modified based on how much better or worse he was than the league average. However, everything is relative to the same context for his opponent. Modifications are made based on how far each player is from the league average.

For example, if a .303 hitter from the National League in 2007 - when players hit .266 - faces a .303 OAV pitcher from the same league and year, the hitter will hit much better than .303 because the batter is far better than the league average and the pitcher is far worse than the league average. Whereas, if a .303 hitter from the National League in 1930 - when players hit .303 - faces a .303 OAV pitcher from the same league and year, the batter will hit .303.

In our SimLeague Baseball player search, we show users how a player would perform against a historically average opponent in our log5 normalization. These are indicated by # and are probably the best indicators of likely sim performance to use when crafting teams.
 

6. Tell us about your weekly MLB Power Rankings.  What type of data is used and how do you derive those rankings?

With the assumption that our simulation engine is the most thorough and accurate way to account for all of the statistical interaction in the game of baseball (hopefully that can be assumed based on the answers above), the MLB Power Rankings are the best possible way to evaluate the 30 active teams in baseball at a given time (which happens to be each Sunday night for the article).

We take the active roster with the stats that each player has accumulated thus far in the season and we simulate each team against every other team 100 times 50 at home and 50 at the opposing ballpark, so that all five pitchers in the current rotation start ten times at each location. Some with limited data from this season have their stats modified a bit to prevent anomalies that could seriously cloud the output.

The MLB Power Rankings are posted in our Beyond the Boxscore section every Monday morning. In late July, our MLB, NFL and College Football Power Rankings will also begin appearing weekly on FOXSports.com.
 

7. Did you run any simulations before the MLB season began to forecast who would win each division?  If so, who did WIS pick?

Actually, we did not. We pride ourselves on our acute attention to detail and our desire to be as accurate as possible. Getting an input in for every possible cup of coffee, injury or other transaction, has been something for which we have previously lacked resources. For 2009, we will have comprehensive player and team previews as well as in-season simulations of the power rankings and stats and standings for the “rest of the season” each week.
 

8. Moving away from the living, talk to us about your online baseball games.  I am personally addicted to Hardball Dynasty, which is by far the best baseball game I have ever played.  Make your sales pitch, son!

It sounds like you just did for me. In all honesty, I completely agree with you. Hardball Dynasty is the best baseball game that I have ever played. Clearly, I would appear to be someone who is supposed to say that, but that does not diminish its truth. Growing up, I played every baseball video and board game that I could – and I still do today. Nothing touches Hardball Dynasty’s depth or level of control.

As an owner of a Hardball Dynasty franchise, users are also the GM and manager(s), in total control of the on and off-the-field play of all minor league teams and the big league team. The goal is to make the best decisions to outwit 31 other owners in the “world” who try to build the best major league ball club that can win and sustain a high level of play.

Build for the future or spend in free agency to win now? Risk a draft choice on guy demanding a very high signing bonus or play it safe? Should you trade a promising 22-year old for a proven veteran? Call up the hot prospect for a playoff chase or wait until he is a little more seasoned? Spend more on coaching and player development or on physical training and medical staff? Scout more in high schools, colleges or internationally? Arbitrate or sign long-term? Protect a 40th man or look for someone in the Rule V draft?

In Hardball Dynasty, you get to answer all of these questions and more.
 

9. In your online games, how closely do you follow the rules, regulations, and nuances of actual MLB?  Like for example, in Hardball Dynasty, I can sign international free agents, Rule 5 draftees,  and deal with signability draft picks.

The goal is to always be as realistic as possible, though we need to consider that this is a game played by those who are not able to do this full-time like their GM or coach/manager counterparts. Hardball Dynasty incorporates facets of the game that are covered by more than ten people working fulltime for a real-life franchise, so we do our best to only really ask our users to make the most important, franchise-altering decisions. The depth of possible decisions and amount of information that can be consumed in the game are unparalleled, but it is not necessarily required of every owner to do so.

To better answer the question, the few minor nuances of the MLB that are not currently in Hardball Dynasty, are likely withheld because we either have not found a way to program them into a simple form that is quick and easy for an owner to address without being exploitable by those who can spend significantly more time on the game than others or because the game, which is just now two years old, has not matured enough to handle the added complexity.
 

10. What is the type of baseball fan which you think would enjoy your games? Just someone who plays fantasy baseball?  Or just the hardcore baseball freak?

We feel that among SimMatchup, SimLeague Baseball and Hardball Dynasty there is plenty for every level of baseball fan to enjoy. SimMatchup caters best to any fan who roots for his or her franchise and wants to see how recent or past teams from that franchise stack up against others.

SimLeague Baseball is very similar to traditional fantasy baseball with regards to constructing a team, except that players from throughout the history of professional baseball are available and that the owner has managerial control over lineups, pitching staffs, etc.

Because of the depth of control and information available, Hardball Dynasty is likely most cherished amongst the “hardcore baseball freak” community, but we have done quite a bit to the game to ensure that it can be played and appreciated by any baseball fan (or anyone interested in becoming a baseball fan).