Turnovers often lead to lost games. The play in any football game that has the greatest chance of leading to a turnover is a punt return. There were 842 turnovers in the regular season last year on 33965 plays from scrimmage (2.47% of all plays resulted in a turnover). There were 2429 punts last season, 1138 of which were fielded by the returner. On those 1138 fielded punts, 87 fumbles occurred (7.64%). So fumbles on fielded punts happened more than three times as often as fumbles on all other plays.
I’m not sure why fumbles were up and long punt returns down last season. Fumbles were up to close to their decade high and long returns were at their nadir. From these numbers it would appear that 2015 was a bad year for punt returners in general. For comparison in 2011, there were 142 punts returned for 20 or more yards compared to the 94 in 2015.
Right now the punt return and kickoff return spots for the 2016 Broncos are wide open. Allow me to present the data which will make the case for the "safest" or "most conservative" punt returner (whoever that might be) over the fastest or most flashy punt returner (capable of getting large chunks of yardage and/or return TDs). The best possible outcome for the return team is a touchdown. Not only are return touchdowns one of the game’s most exciting plays, but they are huge momentum shifters in any game. The problem is that they are quite rare. In 2015 there were 13 punts returned for touchdowns - 1.14%. Any punt returner was almost seven times more likely to fumble on a fielded punt than he was to return it for a TD. There were seven kickoffs returned for touchdowns of the 1080 that were returned - 0.65%.
Okay - but if your punt returner always goes for the fair catch you never gain yardage. That’s true, but we need to look at how often punt returner gained "big" yardage without taking it to the house (>20 yard return). I should also note here that while attempting a fair catch is safe, it still is not without risk as a muffed punt return is still considered a fumble. There were 94 punts returned for >20 yards in 2015 - 8.26%. So punt returners were slightly more likely to get a big return than they were to fumble. While a big punt return can shift field position and swing momentum in your favor, a lost fumble always swings momentum in the other team’s favor. So the possible reward (>20 yard punt return) is still outweighed by the possible risk of the fumble.
Look at it this way - how often does the offense gain 20 or more yards on a play? In 2015 there were 2027 plays that gained >20 yards, 5.96% of all plays from scrimmage. While this number is lower than the 8.26% for punt returns, the risk of a fumble on any given play from scrimmage is still three times lower than on a punt return. So >20 yard gains happen just slightly less often on plays from scrimmage than on punt returns with one third the chance of a fumble. That means it is smart to force your returner to fair catch all returnable punts and allow your offense to do the heavy lifting.
So why don’t all special teams coaches tell their punt returners to fair catch everything? The reason is that there are special players who are great at returning punts (Darren Sproles for example) AND who are also very good at holding onto to the ball. Every punt returner thinks he is Darren Sproles (or Devin Hester in his prime), but few are. There were 24 players who returned 20 or more punts last season - 19 of those 24 fumbled at least once on a punt return. Additionally, attempting a fair catch does not guarantee that a fumble (or muff) will not occur. So while fair catching a punt is safer, it still could lead to a turnover.
The final consideration when selecting a returner has to do with what else that player can do on the team. To use a roster spot for a guy who only does returns (ala Trindon Holliday) is wasteful particularly if that returner is a small guy who isn’t capable of playing on coverage teams. Having your returners as guys who fill other roles on the team (ala Omar Bolden and Emmanuel Sanders) makes much better use limited roster spots. If Khalif Raymond does make the team as the returner (and he looks to be the most dynamic of the potential returners), he will need to fill a role on the offense to offset his lack of ability to be used on coverage special teams. My preference would be to use a "hands" guy at returner who is low risk, but also low reward (Jordan Taylor, maybe). A player like sunshine should also get some use in the offense as the #4 or #5 receiver. I don’t like the idea of using Emmanuel Sanders as the punt returner this year because of the potential for him to get injured while returning a punt.