Keeper rotation

Editor's note (02/16/2013): The combination of DC and Chicago keepers was artificially low because there is a gameweek where neither team plays. Avoid this combination, despite previous advice that it was the best combination to select.

Keeper selection is an interesting issue that has been explored dozens of times for other fantasy soccer games. The general consensus is to find two cheap keepers who rotate having good matchups. I'm a big proponent of this strategy as it allows us to upgrade our squads in other areas. The other strategy is to buy one premium keeper and one cheap keeper and play the premium keeper in the vast majority of gameweeks. In the official MLS game, my gut is telling me to choose the two-cheap-keeper rotation strategy because of the bonus points players receive for recoveries, CBI, etc.

I took the expected goals conceded each week by each team for the first 10 gameweeks and contrasted them with every other team. I then chose the lower value for each gameweek for each combination and used that as my choice keeper for the week. From there I simply added up the value for each of the 10 gameweeks for the following values. The values in red text indicate that there is at least one gameweek where neither team plays.

I also picked out the top 10 combinations for each price combination possible for starting keepers. There were a couple teams (Vancouver, New York, Dallas) where I wasn't sure who would be the starting keeper and would result in different pricing, so I used my best guess for each team. Combinations in yellow indicate at least one missing gameweek.

I should also make note that LA was a completely different team last year with Gonzalez in the lineup. However, even when I adjust the numbers based on their goals conceded with Gonzalez, they still do not have more than one combination in the top 10 overall.

The following keeper combinations have a week where neither play, so their numbers are too low:
Houston/Seattle
Houston/LA
Seattle/LA
Philly/Toronto
Philly/Vancouver
Toronto/Vancouver
Chicago/COlumbus
Chicago/DC (missing 2 gameweeks in fact)
Columbus/DC
Chicago/Dallas
DC/Dallas 

Player projections - defenders

We should approach defenders a bit differently than we approach our attacking options. Since a large part of how defenders score points for us in fantasy is collecting clean sheets, we essentially can just compare which teams are likely to collect clean sheets. Though we are definitely selecting individual players for our defenders, we are also selecting defensive units. If I want to choose Corey Ashe, I am effectively buying into the Houston defensive unit as a whole. Because of this, I tend to choose the cheapest option in each unit who will be a starter unless there is reason to believe one of the more expensive players will produce more bonus points or goals/assists.

Player projections - Midfielders

The midfielder situation isn't nearly as cut and dry as the forward situation. Midfielders tend to get more of both the attacking and defending bonus points in the official MLS game so their scores seem to have a narrower variance for players we are actually considering. There aren't really any clear breaks where there is an obvious way to group the top couple players. The closest I can see happens around Shea Salinas ranking 8th where if we ignore Salinas, we see a drop of about 20 points over the season.

I'm going to leave out Donovan who is off being Sad Landon and we don't know when he will return. So, that leaves us with Rosales, Zusi, Davis, Chavez, Bernier, and Pontius. I also think that Ferreira, DeRo, and LeToux all have potential to give us fantastic returns as they are all coming off tumultuous season of injury or transfers.

Player projections - Forwards

As I mentioned in an earlier post, I created a database with almost all the data from last year's official MLS fantasy game so I could actually make some educated decisions in this year's game. I converted all the categories to a per-minute basis so I could then project player scores as if they played every minute of the season to try and find the best value for money players available based on 2012 stats and 2013 prices. I calculated how many points per game a player would have scored last season if they had played every minute for: goals, assists, clean sheets, CBI, recoveries, crosses, key passes, big chances created, big chances fluffed, and goals allowed.

No, this does not include absolutely every scoring category. And no, my calculations aren't going to be perfect projections - partly because if a player gives exactly one cross per 90 minutes, my model gives them 1/3 point per game, but mostly because this is projecting future returns based off previous returns and does not take into account actual playing time, playing a different role in the team, different team quality in general, etc. There's a million different variables that will make my projections wrong, but I'd rather base my picks on something other than pure speculation and gut feeling. So, here it is:

There's really Higuain, Henry, Gordon, then Wondo, then everyone else. If you don't have one of those four players, you are making a huge mistake. It is reasonable to assume that Gordon won't get enough minutes to be a reliable option. So we are really just deciding between Higuain, Henry, and Wondo for our big-money striker or maybe some of you even want to pick two of them - I see no problem with that. I am going with Wondo to start my season because San Jose are predicted to score the most goals over the first 3, 6, and 10 gameweeks (see my earlier posts for more on that) and they don't have any blank gameweeks to start the season.

Predicted goals allowed - first 10 weeks

I used the exact same methodology to predict goals allowed as I did for goals scored. Again, I want to emphasize that these are 100% based on last year's results. Adjust them as you see fit (I'd recommend rating LA higher than last season, especially with a healthy Gonzalez and upgrade at keeper). Generally, teams playing at home concede fewer goals than they do on the road, so it is a good idea to play defenders playing at home and sit defenders playing away, especially against strong opposition.

Since having a low average goals conceded per game does not necessarily translate to having a high number of clean sheets, we need a different way of predicting who will have clean sheets. I have simply tabulated how many times each team is expected to concede fewer than 1 goal and also fewer than 1.2 goals.

Predicted goals scored - first 10 weeks

The following tables are calculated based off of 2012 goals scored and conceded per game home and away by each team. I simply took the average of goals scored per game by the home team (or away team) and goals allowed by the away team (or home team). This is far from a precise prediction, especially in MLS where so much can change from one season to the next in terms of rosters, player roles, and team playing styles.

This first table shows how many goals each team is expected to score each game for the first 10 weeks. Trying to plan any further out that this is absolutely pointless since we can't know how much each team's quality has changed on either side of the ball. A blank space means they don't have a game that week (or two in the second column of week 8 marked "double"). Teams that had a high-powered offense last season like San Jose will generally look better in these charts. Feel free to adjust based on your own expectations about how a team will do relative to how they fared last season.

This second chart simply adds up how many goals each team is expected to score over a given period of time to start the season. I tend to plan about half my team (generally the more expensive players who I expect will see their price rise) for the 6-week time-frame and the other half for the first couple gameweeks, planning out my first two or three weeks of transfers.

2013 MLS fantasy insight is on the way!

Hello readers!

First, I just want to say that I don't intend to run this as a weekly blog this time around. I simply don't have the time or commitment to do so. I will be doing a preseason breakdown based on last season's statistics.This year I will be looking at the official MLS game instead of the ESPN game.

I managed to input into a database all of the data (such as recoveries, crosses, key passes, etc.) for all players with at least 1,000 minutes played last year. I've separated them out by position and will be using that data to project how many points a player could earn over this season.

I'll also do my typical projection of how many goals a team is projected to score and concede each game based on how many they scored and conceded home and away last season.

So, expect a flurry of posts over the next two or three weeks and then back to blackout for several weeks until I find the gumption to give my revised projections and suggestions.