The optimal use of hook numbers is a Scoresheet nuance that inspires a good deal of debate. Here
is my take on the issue:
Starting Pitchers
Most Scoresheet players use a hook number for our starting pitchers in the 4.5 to 5.5 range. The
consensus of opinion among us Scoresheet-talk members is strongly in favor of low hook numbers such as 3.0 – 3.5 for
your weakest starting pitchers – assuming you have a deep enough bullpen to cover those quick hooks. The logic is that
you minimize use of those starters when they are having a bad week. As is accepted wisdom, relievers tend to have lower ERAs
than starters, anyway, so it is better to make more use of your relievers than your weaker starters. Depending on your draft
philosophy and needs, you hopefully have enough relievers to avoid Pitcher AAA even with such short starter hooks. As I recall,
even the Scoresheet designers themselves admit that using low hook numbers will keep starters’ ERAs down significantly.
Tests have been made to back this up.
Of course, if you bullpen is quite thin, and you are struggling for 5 full starters, then your
hook numbers should be high – such as 7.5.
Relievers
What is more hotly disputed is how low relievers’ hook numbers should be. I believe most
Scoresheet players use hook numbers for relievers in the 1.5 to 2.5 range.
Contrary to his advice on starting pitchers, SWARP baseball stat forecaster Ken Warren strongly
advocates high hook numbers for his relievers (2.5 - 4.0). His argument centers on his contention that when a reliever comes
into the game with runners on base, there is about a 33% greater chance those runners will score than when no pitching change
is made. This is due to the ERA matching controls getting turned off. Because relievers pitch so few innings in a week, they
often have an ERA of 0.00. Even if they have given up some runs already, you never know if their ERA for the week has been
matched, so taking him out might be hurting you more than helping. In either case, there will be a strong pull against them
giving up runs. Hence, with ERA matching turned on, you are protected against
their runners scoring. Without it, you are not. Even if the new pitcher’s ERA for the week was 0.00, there is nothing
stopping the Scoresheet simulator from giving the upcoming batters their share of hits and walks until it is the new pitcher’s
runners’ turn to score. In the meantime, they may have cleared the bases. This is further reason any hook below 2.0
is not considered optimal, because the pitcher could load the bases without giving up a run, he gets yanked and instead of
your reliever working himself out of a jam due to his ERA matching, a new reliever comes in and gives up his normal amount
of hits and walks while the runners come marching home.
Why is a low hook number good for starters, but not relievers? 1) Starters pitch more innings,
so when you hook a bad one early you are getting rid of more bad innings. 2) Starters are generally only used once, so he
won’t come back to ruin another game with unused bad innings. RotoWire contributor
John R Mayne reminded me in SS-talk that when a top reliever is hooked early during a bad week, he is more apt to keep coming back and messing up your games.
One of Ken’s Scoresheet-talk responses to my prodding about the reasoning behind his methods gives a rule of thumb for setting up hook numbers:
Well
every team is different, but for a
strong
team with a good bullpen the
following
strategy should work out the best.
Top starters - 6.5 - 7.0
Average starters - 4.5 - 5.0
Poor starters and relievers
who start - 3.0
Top relievers - 3.5 to 4.0
Long relievers - 2.5 to 3.0
For sixteen years, my teams have been as successful as possible using reliever hooks in the 1.0
to 2.0 range. Because relievers pitch so few innings each week, it seemed too much of a gamble to give them high hooks. However,
I was open-minded about this. Using the higher hook numbers recently has worked out well as I wrote a few years ago:
My top relievers did improve with the increased hook numbers. I selected my two most relied upon
relievers’ ERAs and Innings and compared those stats to their real life performances, then compared those differences
this year (’07) to years past when I used lower hook numbers. Using only two relievers reduced the use of relievers
who were not heavily replied upon all year, which reduces the number of innings lost due to lack of prominence in the bullpen
rather than innings lost due to the size of the hook numbers.
I recorded the differences between both of my teams’ my top relievers’ Scoresheet
results and real life results going back to 2001. Going over almost all my line-up changes during that time, in 2007 my average
hook number for the top two relievers on each team was 2.3. On the 2001-2006 teams it was a mere 1.1. My top relievers on
my 2007 teams had lower Scoresheet ERAs than their real life ERAs by an average of 0.50, while the top two relievers on both
teams from 2001 to 2006 averaged no increase or decrease in ERA. (I’ve always been a fan of drafting players with good
defense, which nullifies the all-star affect on my pitching staff. A liberal use of defensive replacements particularly helps
my relievers.) I did not have a greater than usual impact of defensive replacements in 2007 than in the other seasons. Even
the amount of innings improved slightly. Instead of gaining a mere inning over the course of the year, my ’07 relievers
gained 2.3 innings. Here is the spreadsheet for this analysis.
Admittedly, this was a small sample size. I will continue to keep track, perhaps, with some 3.0
or 4.0 hooks!
However, I think we can refine this a little. Ken does not distinguish between long relievers
and middle relievers. In a typical week, my middle relievers and long relievers only get used once or twice. With the deep
bullpens that Ken favors, one can afford to put a low hook number on one’s middle relievers in order to minimize his
bad weeks just as you would a starting pitcher. After all, there is a good chance you won’t have to use him again. Save,
perhaps, just your last righty and last lefty for a large hook number in order to avoid a complete drain on your relievers
or AAA pitchers.
Left-handed fractions
Do you have a left hander who only pitches 50 innings a year? Save him for multiple games by using
0.5 as his hook.
Have no lefties among your top relievers? Use a fraction not divisible by 0.5 such as 2.8. That
way the sim won’t pause to see if it has a favorable lefty/righty match-up mid inning the way it would after two runs
and either one runner (2.5) or two (3.0) runners on base.
Note, however, with higher hook numbers the need for a lefty specialist is a bit less. It would
still be nice to have a couple lefties on hand to relieve your starters. Remember, pinch hitters aren’t used at all
in the A.L. until the 7th inning or 6th inning with two runners on. So, if you bring in a new pitcher
early in the game throwing from the opposite side of the mound as your starter, he will have a favorable line-up to face.
Two Closers?
One radical idea is using two closers. Perhaps, that could be to some advantage in the play-offs,
but I have had seasons where my one closer wasn’t getting enough innings. On the average, I have lost 10 innings a season
during my first four years of owning Mariano Rivera generally using an Earliest Inning of 8. However, if anyone wants to try
it, Ken Warren suggests using a much lower hook number. From Scoresheet-talk Sep. 11, 2007:
I
always have my closer come into the game in the 7th inning. This
makes sure that I always get all their innings and also
makes using
two closers a viable strategy.
The big advantage of using two closers is that you don't waste so
much
of your top pitching in blow out victories. Any pitcher listed as a
set-up guy but not a closer will come into a game no
matter how many
runs you are ahead.
With two closers really low hook numbers are best, although I would
never
go below 2.0. A hook of 1.5 would mean that your pitcher who
pitched a major league shutout in real life would come out
with the
bases loaded and no runs allowed, when in fact Scoresheet would allow
him to pitch ten shutout innings for
you if you hadn't hooked him at 1.5.
Earliest Inning
The more interesting point here is bringing his closer in the 7th
inning. The conventional wisdom among ss-talk and mcscoresheet experts is to always use an EI of 6. In fact, I try to have my first set-up man listed with an Earliest Inning of 7. My second
righty generally has a 6 along with my only lefty set-up (if that is the configuration – ideal, but not always a match
with the relievers in my hand). I agree with Ken that using an EI of 7 is an efficient use of your short relievers as they
never pitch more than 3 innings at a time. Obviously, if you bring your reliever in the 6th inning, he has to be
relieved at some point (unless you are losing and the visiting team). Hence an EI of 7 avoids wasting a relief appearance,
which not only costs an extra inning of available use, but increases the chances of introducing a pitcher into the game who
may have had a terrible week.
The latter reason is why I see a disadvantage in using a closer at
all. It is in closing situations that you least want to risk introducing a pitcher going badly. Yet, Scoresheet will take
out your set-up man for a closer no matter how well he is pitching.
Typical Pitching set-up (for Scoresheetwiz
now that he no longer owns Mariano Rivera)
Starting Rotation Hook
Pitching Ace 5.0 (higher if short on pitchers)
Reallygood Two 5.0
Third Starter 4.0
Good Fourthguy 4.0
Fifth Starter 3.0
General Bullpen Hook Ing. vs.R vs.L
Alt Rightystarter 5.0 1 6 6 (or higher)
Alt Leftystarter 5.0
1 7 5
(or lower)
Reliever Five 1.0 1 5
9
Reliever Four 1.0 1 4
8
Reliever Three 1.0
3 3 7
Reliever Loogy 0.5
3 9 2
Reliever Two 3.5 6 2
4
Top Lefty 3.5 6 8 1
Top Reliever 3.5 7 1
3 (or 2)
