Tuesday, May 28, 2013

Goals scored and clean sheets: comparing historical averages to the season so far

The trend of scoring well below the historical average finally ended. Through the first 10 weeks of the season, defenders were king and forwards were more or less ignored. Many players opted to play 5-4-1 or 4-5-1 formations and lamented the fact that defenders were scoring far too many points relative to forwards. Three weeks ago, that all changed.

There were 2.86 goals scored per game over the last three weeks. (If we scored at that rate over all games this season, there would be 44 more goals scored this season than what have actually been scored). That brings the season average up to 2.48 GPG, still below the historical average. Over the last four years, we saw 2.57 GPG. If we simply apply the historical GPG to this season’s games played, we would have 303 goals - 10 more than we experienced. That isn’t too large to suggest any huge shift for our strategy. However, if we ignore the last three weeks and just look at how far behind we were during the first 10 weeks, we really see why I think it is time to divest from our defenders.

Over the first 10 weeks of the season, there were a measly 2.33 GPG (193 goals in 83 games). If we apply the historical rate of 2.57 GPG to those 83 games, we would have had 213 goals – 20 more than what we saw. That’s 100 points from goals alone and about 150 if you include assists at the historical rate. Most of those points would go to forwards. We’ll assume that about 60% (90 points) of them will go to forwards, 30% to midfielders, and 10% for defenders.

A mere 150 points doesn’t seem like it would make much a difference… and it really isn’t a colossal change. The big change comes when we look at the clean sheets. For this portion, I will assume teams use 4 defenders and a goalie, for 5 total “defenders”. 

So far this season, defenses have kept an astonishing 26.7% of possible clean sheets. Incredible. Through the first 10 weeks of the season, it was an even better 28.3% of possible clean sheets. Over the last four years, the rate of clean sheets is 13.7%. That means that teams so far this season are keeping clean sheets about twice as often as they have over the last four years. If we apply the historical rate of clean sheets to this season, there would be 32 fewer clean sheets this season than we’ve seen in reality (63). 32 clean sheets, times 4 points for a clean sheet, times 5 defenders = 640 more points earned by defenders this year from clean sheets than what would be expected based on historical averages. And just for fun we’ll add in the points lost (-1) for conceding two goals for each of the 20 more goals we would expect from historical rates and we get another 50 points that defenders have scored this year that they wouldn’t have. That is a total of 690 points defenders have earned this season above what is normal. That is HUGE and equates to 7.26 more points for each starting defender/keeper than what they were expected to earn.

Now I’ll apply these expected average adjustments from above (+2.37 for forwards and -7.06 for defenders) to the average score of the top 10 and top 25 for each position.

What to make from all of this? Defenders have been vastly outperforming historical averages; forwards have been slightly underperforming historical averages; and if we apply those averages to what has actually happened, top defenders should still be expected to outperform top forwards, but not by nearly as much as they have been this season. The assumption that those extra goals would be evenly distributed amongst all starting forwards is a bad assumption, because we would expect guys like Wondo, Henry, and Higuain to outscore guys like Tristen Bowen and Lionard Pajoy.

Monday, May 20, 2013

Projections week 13-17

 Projected goals against:
If you missed the explaination for these calculations the first time around, I basically just take the average of the team's goals conceded and their opponent's goals scored based on who is home and away. Those highlighted in red are the gameweeks where the team is predicted to concede less than one goal and are high chance of a clean sheet.

 Projected goals for:
Basically, the same calculation as above, only calculated for the team's goals scored. The cells highlighted in red are games where the team is projected to score more than 1.5 goals.

 Total goals scored during 5-week span:

Goals for/against home/away

We are now at a point in the season where every team has played at least 5 games at home at 5 away. That is the point where I'm going to start using data based only on this season to project goals scored and conceded. I'll share the next 5-week projections in another post, but this one will just show some raw data for each team on how many goals they score and concede per game both home and away. I've ranked them from best to worst:

Thursday, May 16, 2013

Week 15: fail to prepare, prepare to fail.

The double-gameweek bacchanalia of week 11 is over and our rosters are locked for the morning-after bloody mary of week 12’s Philly/LA double. Weeks 13 and 14 aren’t anything special but weeks 15 and 16 loom on the horizon. If you happen to still have your wildcard, use it between weeks 15 and 16 and take advantage of good matchups between now and then. For the rest of us that used our wildcard early (week 3 for me), we need to start planning NOW. Week 15 has 9 teams without a game and week 16 has 5 teams without a game and 4 teams with a double. Oh yeah, then there's a bunch of players gone for international duty to boot.

Week 15 players:
·         Chicago (vs. Portland)
·         Columbus (at Philly)
·         DC (at New England)
·         LA (at RSL)
·         New England (vs. DC)
·         Philly (vs. Columbus)
·         Portland (at Chicago)
·         RSL (vs. LA)
·         Seattle (vs. Vancouver)
·         Vancouver (at Seattle)

Week 16 feast:
·         Colorado (vs. San Jose and at Chicago)
·         Montreal (at Columbus and vs. Houston)
·         Portland (vs. Dallas and at LA)
·         Vancouver (vs. New England and vs. Chivas)

Week 16 famine:
·         New York
·         Philly
·         RSL
·         Seattle
·         SKC

So, what teams do we target to bring in and when do we do it?

Targets to transfer in before week 13 deadline:
RSL and New England – both teams have 3 home games in a row in weeks 13, 14, and 15. I like New England’s defense as the best target for transfers this week as they are projected to keep 2 clean sheets out of 3 based on this season’s goals scored/conceded home and away for them and their opponents. I plan to double-up on New England defenders. RSL’s prospects on defense don’t look as good, with their only projected clean sheet coming in week 13. On the other side of the ball, RSL players are more attractive and are projected to score more goals than anyone over the coming 3 weeks (tied with LA and Columbus). RSL do have a blank in week 16, so limit yourself to a maximum of 2 RSL players. If you don’t like either RSL or the Revs, LA is where you want to invest as they don’t have a blank week during this stretch and project decent enough returns.

Targets to transfer in before week 14 deadline:
RSL and New England players are still fine targets here. However, the best targets at this point are Columbus attackers. Columbus have two home games surrounding an away trip to face a leaky Philly defense. Seattle are projected to score the most goals in weeks 14 and 15 combined and are projected to keep a clean sheet in week 15 so they are worth a look - beware of their blank in week 16. Aside from that, there really aren’t many favorable matchups. You could look at Vancouver who are the only team who play 4 games during weeks 14-16 (two home games in week 16 look tempting). My plan for week 14 is to make just a single transfer and roll the other over into week 15. There aren’t many players I’m excited about in week 14 who play in week 15, and it also allows us the flexibility to shift money around in week 15 and helps avoid potential injury/suspension issues.

Targets to transfer in before week 15 deadline:
At this point, a lot depends on who you have and how you’re set up for week 16. Portland and Vancouver are the only two teams to play in week 15 and have a double in week 16. Vancouver does have a blank in week 17, so I’m not crazy about them as transfer targets after this week. Portland and Colorado are the teams with doubles in 16 and no blank in 17, so Portland players are where I’d look to first – might even triple up on Timbers by week 16. Columbus is projected to score the most goals during this week and don’t have a blank in 16 or 17 – not the most exciting picks going forward, but good enough to fill some roster slots if needed for week 15. If you want to start building again for after week 15, then LA are your guys as they have no blanks and project well going forward.

Top individual transfer targets by week:
Week 13: Morales, Goncalves, Schuler, Farrell, Shuttleworth, Donovan, Keane
Week 14: Higuain, Rosales, Morales, Goncalves, Farrell, Shuttleworth, Oduro
Week 15: Wallace, Valeri, W. Johnson, Higuain, Oduro, Donovan, Keane

My plan:
Week 13: Ashe to Farrell and O. Gonzalez to Goncalves
Week 14: R. Johnson to Oduro
Week 15: Holgersson to Wallace, Hedges to Lee, save third transfer to bring in a player with a double week 16

My spring coaching season officially ends this weekend, so expect a couple posts over the next couple weeks as I will suddenly have free-time on my hands.