Goals per game

Thursday, February 27, 2014

Under the Radar or Over-Owned



UNDER THE RADAR OR OVER-HYPED

Alright, so I’ve published a giant pile of data with a small bit of written analysis to accompany it. I’ve recommended the reddit forum for discussion about strategies and individual players. I’ve revealed my current lineup. Now it’s time to get into some risky picks that I think are flying under the radar and which players are way over-hyped or over-owned.

Under the radar

San Jose defenders – I understand that they are under-owned partially because they have a bye to start the season and have champions league to congest their schedule early on. However, their defenders haven’t even been in any discussion that I’ve seen anywhere. Last year the average team kept a clean sheet 28% of the time. San Jose under Mark Watson (and after the addition of Clarence Goodson and Jordan Stewart) kept a clean sheet 42% of their games – only Portland had a better rate for the full season last year. Specifically, Victor Bernardez was in the top 20% of defenders in both CBI/minute and recoveries/minute – and he can chip in an odd goal or assist.

Matt Hedges – He is in the top 20% for both CBI and recoveries per minute. He can score goals. Dallas’ schedule to start the season is fantastic. Sign me up!

Chad Marshall – Same reasoning as Hedges.

David Horst – It looks like he is winning the battle for the other starting center back role in Houston. Assuming he does, he is insane value.

Thomas McNamara – He has been starting for Chivas in preseason and seems locked in to start. Brilliant option as your last sub to free up some money to upgrade elsewhere.

Mauro Diaz – He is the creator in Dallas. His PP90 and underlying stats last season were insane and if he continues to create close to that same rate, he will be some of the best value in the game.

Dillon Serna – Same reasoning as McNamara.

Gaston Fernandez – He will start for Portland. Depending on Valeri’s health, it might be as CAM or it might be as LM. Either way, he is a great combination player and has been on set pieces so he’ll collect bonus points. Plus he’s reasonably priced at $8.0.

Stefan Frei – He will be the starter in Seattle. He is cheap. Seattle’s defense has been solid in preseason.


OVER-OWNED, OVER-HYPED

Defenders

Deandre Yedlin, 26.9% owned – Outside backs just aren’t as good as center backs for consistency and their ceiling is lower. After clean sheets, the strongest correlation for defenders and PP90 were goals scored and CBI… things outside backs just don’t get. Yedlin will get forward plenty and notch some assists, but they will be sporadic and he won’t get many bonus points (as is true for outside backs in general). I’m not opposed to selecting outside backs, but if I do they have to be cheap. Yedlin isn’t. There are plenty of centerbacks with higher fantasy ceilings than Yedlin at similar prices or cheaper.

Dylan Remick, 23.9% owned – All this hype was because he looked to be starting at left back for a couple games at the start of the season. Leo Gonzalez started for the Sounders in yesterday’s scrimmage and is well on his way to recovery. Remick might not even start in week 1 and could see a mass exodus and is a massive risk of price drop.

Michael Harrington, 15.5% owned – Same reasoning as Yedlin.

Sean Franklin, 9.8% owned – Same reasoning as Yedlin.

Omar Gonzalez, 7.2% owned, He’s a great player and a great fantasy asset. His schedule just sucks to start the season. It is so bad that he’s not in my side.

Seth Sinovic, 7.0% owned – Same reasoning as Yedlin.

Midfielders

Juninho, 17.7% owned - Defensive midfielders suffer for the same reasoning as outside backs relative to others within their position classification. Midfielder PP90 is most correlated with goals, assists, key passes, crosses, and big chances created. None of which are racked up by any defensive midfielders.

Adam Moffat, 15.7% owned – He’s not a starter and is a waste of money. Either downgrade him to a 4.0 player or upgrade him to someone who will start.

Jimenez, 14.2% owned – Same reasoning as Moffat.

Rodney Wallace, 13.3% owned – He’s still recovering from injury and won’t start the season in the first 11. He might not even gain back his starter role as Portland brought in Steve Zakuani (also injured) and Gaston Fernandez, who shifted wide left earlier this week when Valeri subbed on.

Michael Bradley, 11.8% owned – He’s new to the league and his entire team got retooled in the offseason. Strike one. He’s a deep-lying midfielder. Strike two. He costs $10.0m. Strike three.

Ben Zemanski, 8.4% owned – Same reasoning as Moffat.

Forwards

Mike Magee, 25.8% owned – I demand a lot of my forwards. They have to either be cheap (Plata, Saad, Urruti) or be a creative force for their team to collect bonus points (Keane, Henry, Higuain). Magee scores goals, which is nice. I don’t see him repeating a 20+ goal season this time around and his bonus point generation is lacking.

Robbie Keane, 17.1% owned – Same reasoning as Gonzalez.

Jack McBean, 12.0% owned – Same reasoning as Moffat.

Connor Casey, 9.7% owned – Same reasoning as Moffat.

Jack McInerney, 7.8% owned – Same reasoning as Magee.

Monday, February 24, 2014

PP90 and finding undervalued players



I found some time to actually work with some data and do some analysis! This post will combine two ideas: PP90 and finding players who were underpriced based on last season’s production. PP90 is simply a player’s total points divided by the total minutes they played last season. I adjusted for players who had their position reclassified (e.g. removed clean sheet points for Mike Magee) and also added back points from Big Chances Fluffed. The data does not adjust for second assists, as I just don’t have time to manually enter that data (Valeri, who had the most second assists last season, saw his PP90 increase by less than 5%). All data is from the regular season, as I feel playoffs are such a crapshoot as teams play differently in the do-or-die situation.

Looking at undervalued players is the bulk of this post, as I believe it is more important than looking at raw PP90 because of salary cap restrictions. I ran a regression that looked to see what PP90 a player with a given price is expected to get (e.g a player priced at 11.0 should get about 6.25 PP90). The column labeled “difference between expected and actual” is the important one for these purposes. The larger the number, the more undervalued the player is.

There’s a catch. PP90 assumes that the player played every minute and that their rate of return remained the same. We know this is absolutely untrue for a few reasons: players get subbed in/out so don’t play 90 minutes, and maintaining a constant rate of return is basically so improbable that we know that assumption is wrong. However, this is the best way to compare players because it normalizes our measure of comparison so we are comparing apples to apples. We must couple this knowledge with our knowledge of who is starting (Keane), who is a backup that will get starts when someone else is injured or away on national team duty (Opara), and who comes on (Alhassan) and off (Rosales) as substitutes.

If a cell is highlighted in red, it is within the top 10% for all players for that category. So, Robbie Keane is in the top 10% for both PP90 and how undervalued he is, whereas Diego Valeri is only in the top 10% for PP90.

FIRST TIER – priced 10.5 to 11.0

This is where we want to start to build our team; the big guns who will often wear the captain’s armband. Even though these players are the most expensive in the game, many are still underpriced. Robbie Keane and Landon Donovan lead the charge, but shouldn’t lead many lines to start the season because LA have a couple bye weeks and a bunch of away games to start the season. 

Those of you who followed my blog last year know how much I value the consistency of bonus points over the more sporadic points from goals and assists. I recommended Higuain over and over again in the lead up to and throughout the season and will do so again, though he is more appropriately priced this year. Valeri and Zusi are the other players I highly recommend in the top tier. I’ve also had Brad Davis in and out of my lineup to start the season, mostly because of Houston’s favorable schedule to start the season.

SECOND TIER – priced 9.5 to 10.0

This is an interesting tier to pick from. A lot of the players have big reputations and are better players in real life than they are in the fantasy game. Olave, Collin, and Martins are the three players from this tier that I’ve considered for my team. I’ve gone with Martins for my team due to Seattle’s favorable schedule. Wondolowski is another interesting player because last season’s fantasy output was hampered by him playing on a broken foot most of the year. He’s a bit of a poacher and doesn’t produce a ton of bonus points, but I think he’s not as over-valued as these stats suggest.

THIRD TIER – priced 8.5 to 9.0
 There’s a lot of interesting options in this tier and most are pretty appropriate priced. Even with Wallace losing his clean sheet points because of his position reclassification, he is way undervalued. He is injured to start the season and Portland’s signings might see him left out of the 11 even if he were healthy. Le Toux and Futty are also possible victims of reworked rosters. I’m going with a strong backline this year with centerbacks who score goals and rack up CBIs. I have Chad Marshall and Matt Hedges in my team at the moment while Okugo and Bernardez have been in and out. I also am looking closely at Kenny Miller as I think he’ll play a bigger role with Camilo gone.

FOURTH TIER – priced 7.5 to 8.0

This is the tier you will either ignore completely or have a bunch of players on your roster depending on the general strategy you take: a bunch of solid players or several players from the first two tiers and a bunch of players from fifth tier or below. This is where you want to take a risk or two, in my opinion and many of these players have the potential for much higher returns than they saw last year. Salinas is one of the most underrated players in the league and his style of play is conducive to fantasy… if he can stay healthy. Lots of these players have changed teams (Rosales, Hurtado, Cooper, Beitashour) or have their minutes in danger due to new signings (Traore, Michel, Casey, Neagle). I like to avoid such situations until we can see who is the first choice and how their role looks to contribute in fantasy terms.

FIFTH TIER – priced 6.5 to 7.0

The budget starters. Don’t really want to start most of these guys, but don’t really have a choice. There’s a lot of potential value to be had in this group, if these guys can lock down minutes. Ike Opara is the obvious standout, but he isn’t a starter until Besler goes to the national team (at which point I’ll grab him). Barklage, Saad, Plata, Osorio, Duka all look like great plays this year if they get some serious minutes. Of those, Duka looks the most nailed on and is in my team. Saad is also in my team at the moment as my one risky player who might not start.

I want to make a point or two while discussing this tier, as most of these players are outside backs or defensive midfielders. Avoid them both. Defensive midfielders are almost always overvalued in the fantasy game and their ceiling simply isn’t high enough to warrant inclusion. 2-4 points is a pretty normal return for these guys, which doesn’t cut it for me. Outside backs also aren’t worth it when compared to center backs. Aside from clean sheets, goals and CBI had the highest correlation with PP90 for defenders, which are the categories center backs score higher in than outside backs.

SIXTH TIER – priced 6.0 and below

I did not include these players in this analysis because there are very few starters within this group. That said, there are always players in this group who will be starters for chunks of the season. A good way to keep up with these potential bargain basement finds is to read and participate on reddit.

Monday, February 10, 2014

Why looking at just overall points from last season is a terrible strategy.



Every year around this time I see people getting into the discussions of player X over player Y. People try to use stats to back up their claims (rightly so) but don’t understand why some stats are better and more valuable than others. So, I’m just going to bluntly say this: total points is a garbage stat to use.
Let me make the disclaimer that this article is aimed at people who will take an active managing their teams and are willing to look a bit deeper at some stats. Why? Because we need to look at multiple stats and factors about why a player is worth owning. 

Here’s the first three questions we need to answer to start narrowing down our list of potential players for our fantasy teams.

1.       Is this player a starter?
2.       What was this player’s PP90 or PPG?
3.       Did this player change positions in real life or as classified in the fantasy game?

“Is the player a starter?” is the most important question to ask. If the player isn’t a starter, they won’t be earning points. Not only do they miss out on the basic points from just playing in a game or at least 60 minutes, but they are also missing out on opportunities to score points in any other way that requires them to actually be playing. Last year I took some risks with my lineup to start the season and ended up with a couple players who ended up not being starters, had weak scores the first couple weeks, and was forced to use my wildcard way too early in the season.

“What was this player’s PP90 or PPG?” is the meat and potatoes of player analysis. Why is this better than looking at a player’s total points? Because it takes into account time lost due to international duty and injuries (yes I know that some players have perpetual injury issues). As an example, we’ll look at similarly priced Austin Berry (144 points, 3060 minutes played) and Jermaine Taylor (120 points, 2142 minutes played). Yes, Austin Berry earned 22 points more, but he played almost 1,000 minutes more over the season. If we just look at last year’s regular season PPG, Austin Berry scored 0.4 fewer PPG. If we extrapolate that over the season, it is 13 points lower score than what Jermaine Taylor would have scored (assuming his production per game stayed the same). If we do that same analysis while looking at PP90 instead of PPG, the difference is twice as large. 

Now, we have to pay attention to which measure to use. Some players almost always are subbed out of the game, so it doesn’t make sense to use PP90 for those players. Last year, Mauro Rosales was a perfect example of this as his PP90 was a whopping 1.7 points higher than his PPG. However, I don’t think he lasted a full 90 minutes in any game all season long. So, we know that it makes more sense to utilize his PPG instead of his PP90.

“Did this player change positions in real life or as classified in the fantasy game?” is another important question to ask and the answer can either hurt or help a player’s fantasy production. Amobi Okugo was classified as a midfielder last year but actually played as a center back. Because of that, he missed out on 33 points he would have earned if classified as a defender, as he is this year. On the flip side, Rodney Wallace was classified as a defender but played as a midfielder and earned 30 clean sheet points that he won’t have the potential to earn this year. These position reclassifications (as well as changes to the scoring system) are taken into account in my upcoming tables that show PP90. There are also cases of players who will play a different position this year than they did last year. Landon Donovan looks like he will be one of them, as The Bruce said that Donovan will likely be pushed back into midfield this year. It is impossible to quantify how that might change his fantasy production, but I’m willing to bet it will be worse this year than last year, simply because of the actual position change.

Beginner's guide to fantasy MLS



The old beginner’s guide is a couple years old now and sorely out of date, especially with the rules changes this season. So I’ve updated bits and pieces of it this year to make it applicable to this year’s game.

Rule 1: Read and understand the rules of the game.
This cannot be overstated. The quickest way to become a decent fantasy manager is to simply understand how your team will score points. Understand the scoring, understand the deadlines, understand how substitutions come into play, understand how the captain’s armband works.

Rule 2: Don’t forget about your team.
I see it happen all the time; people select their team and then just leave it alone. You will not win if you do this. Injuries happen, players fall out of favor, the national team comes calling, your star player’s team doesn’t have a game this week and you left the captain’s armband on him. These are all things that can kill your chances of doing well in fantasy. You have to pay at least a little bit of attention to your team so you can make appropriate substitutions and transfers. You can always check the upcoming fixtures table over at MLSfantasyboss to make sure you’re staying on top of that.

There are a couple of websites that track injuries that I check the night before deadline. Neither have been updated for this season, but are valuable resources once the season kicks off.

http://sports.yahoo.com/mls/injuries

http://www.mlssoccer.com/mls-injury-report

Rule 3: Don’t choose defensive midfielders.
Defensive midfielders are a complete waste of space in your fantasy roster. Players like Alonso, Beckerman, Larentowicz, Juninho, etc. are all important players for their team in real life. None of them will ever find a spot in my fantasy team because their primary role is to win the ball, not score and set up goals.

Rule 4: Don’t fall for early-season over-performance.
There are always a few players who score a goal/assist or two in the first couple games and everyone jumps on board and transfers them in. Sometimes this is the right move, most of the time it’s not. If they weren’t on our radar before the season, chances are their early returns will not continue. Last year had a prime example of this: Jack Mcinerney came out of the gates strong racking up points and managers transferring him in. Then he completely disappeared and did nothing the second half of the season, but a lot of managers were still playing him.

This rule has a bit more leeway than the first two, as sometimes it might be a good idea to bring in a forward or attacking midfielder early if it looks like they will explode right out of the gate.

Rule 5: Pick the known quantities
This one seems obvious and most people have it down. If you’re new to MLS and know a player’s name, there’s probably a good reason for that. Players like Zusi, Keane, and Valeri are pretty safe bets that anyone should feel comfortable enough picking them.

Rule 6: Don’t pick a bunch of unknown quantities.
This is essentially the inverse of rule 5. Every season there are a few big-name transfers into MLS and a crop of rookies that get a lot of hype and people get too excited and pick those players for their fantasy team. Often, first-time MLS players take a bit of time to get used to the style of play and gel with their team. Even though a player looks like a quality signing on paper, they tend to take a bit of time to adjust to their new teammates and level of physicality of MLS.

Rule 7: Pay attention to a player’s actual position.
There are always times where the game lists a player in a position they aren’t actually playing in. Sometimes this works in our favor, and sometimes it doesn’t. Players playing in a more advanced position than they are listed as can add opportunities to score offensive points but still collect clean sheet points from the positions they are listed as. Last year’s examples are Rodney Wallace (listed as DEF but really played as a winger) and Mike Magee (listed as MID but really played as a FWD). Time will tell on this year’s out of position players, but a speculative list can be found here.

Rule 8: Don’t wear your team colors.
Aside from rule 2, this is probably the most common mistake I see fantasy managers make. If you are a fan of San Jose, you can’t just pick 4 San Jose players while completely ignoring everyone wearing an LA Galaxy shirt. You’ll be missing out on a lot of really good players from LA and might be letting your fandom blind you to how bad Steven Lenhart really is for fantasy soccer.

Rule 9: Players generally perform better at home.
It is a well-known fact that teams play differently based on whether they are home or away. Away teams play more defensively, which oddly means that the home team is more likely to collect a clean sheet and also score goals. Keep this in mind when selecting your squad. You obviously want to start your superstars, but your midrange and cheap options should take into account whether they are home or away this week.

Rule 10: Use your transfers wisely.
We are allotted a ton of transfers in the MLS fantasy game, 2 per week if you average it out. It generally is a good idea to look ahead a few weeks when deciding to bring in a player to make sure that the upcoming fixtures are favorable and he will be likely to produce fantasy returns.

However, sometimes there is a single fixture that is just too good to pass up. With so many transfers allotted, it probably isn’t a bad idea to bring in a Keane, Valeri, or Higuain for a single game if they are playing at home against a team like Chivas who concede a lot of goals on the road.