Explanation and Overview of Stats Displayed on
this Site
On Probullstats, we use and refer to a great many statistics which
do not exist outside of here. This is a guide to what those are
and what they mean.
Almost every stat we compile or examine here COULD be compiled
or examined somewhat differently. giving somewhat different results.
Rankings for riders, bulls, and events could be made to look different
in an infinite number of ways. Statistics are often used to promote
some idea or policy, such as in politics, and there are an endless
number of ways in which a statistician may work numbers to a specific
agenda. For example, I could make a convincing argument backed by
the data on hand here at probullstats for at least 8 different bulls
to be the number one ranked bull. That amounts to building a case
for a specific result. That is not what I try to do. I start with
a question - like... what are the ten best bulls? Who are the most
efficient riders? And I try to answer these questions as honestly
as possible. I don't wish to arrive at any particular answer, just
the most accurate one possible given the data we have. Naturally,
there are questions that our data can never answer. What bull is
the greatest of all time? We don't have enough data to answer that.
The three main areas of inquiry here are riders, bulls, and events.
What are the best or worst bulls? Who are the best and worst riders?
What are the best or worst events in our data set?
BULLS:
As explained elsewhere, bulls are summarily
ranked on the basis of their average marking. It's the best data
we have for this purpose. It is not perfect, and the judges surely
do not always get it exactly right. But over time, it is the best
indicator of the relative quality of any bull.
I keep two ranking tables that are recompiled each time new data
is added to probullstats. I keep a table of the best proven bulls
in the entire data set, and a table of the best proven bulls that
are still actively competing - so far as we know. By "proven
bulls" I mean bulls that have 5 or more marked in the database.
Usually, once a year, I find all the bulls that have not had an
out in over 12 months, and mark them as inactive. In both these
tables, I store each bull's average marking overall, and several
other relevant stats including Probullstats
Power Rating - which is explained below
These tables allow us to display a number of rankings, including:
The top 100 bulls with 10 or
more outs, the Top 200
Active Bulls, and much more. Most of the available stats for
bulls can be found on our bull rankings
page.
Round Wins and Winning Percentage
For Bulls
On most bull profiles mid-page on the right, you will see a line
like this one:
Go Round Appearances: 60
Round Winning Outs: 21 (.350)
Bull Riders win money and points by winning events (or more specifically,
by winning go rounds at events). This is the equivalent measure
for bulls - showing how often they win at the go round level. Go
Round Appearances is simply the number of go rounds the bull has
appeared in. Round Winning Outs is the number of times the bull
has recorded (or tied for) the high marked out in a go round. The
number in parentheses is just a batting average - or more appropriately,
winning percentage. Wins / Appearances. There's a slight
difference from the straight batting average scenario, though, in
that a bull could have two round winning outs at the same round
by being out more than once. (e.g. two day events with a short round
both days).
This stat also says something about the quality of bulls that usually
appear with any particular bull. You'll notice that many great BFTS
bulls don't have a high winning percentage, because it's very hard
to have an round winning out at a BFTS event. But at smaller events
where the bull quality level is not so high, winning percentage
will tell you to what extent a bull tends to be the best in his
own herd, or among the bulls he usually appears with.
The numbers above - .350 winning percentage - belong to C21 White
Magic as of late April 2008. So you can see that it takes quite
a dominant bull to win .350.
From a bull rider's perspective, you can't draw any better in any
go round than drawing the bull that records the high marked out
in that round. The winning % stat shows how often that happens with
each particular bull.
Power Ratings
There are two kinds of Power Ratings displayed here. The PBR
Power Rating for PBR bulls - that you see on the PBR telecasts,
and in each bull's PBR box stats is:
Average Marking from the past 12 months (magnified to 100 point
scale)
PLUS
Buckoff % from the past 12 months
DIVIDED BY 2
Probullstats Power Rating for bulls is a stat
that I've played around with, and I've tweaked it nine ways from
Sunday over a long period of time before including it among the
bull rankings. My reason for fooling with it at all is that in the
course of looking at several years worth of NFR go rounds and short
go rounds, final rounds etc. at major events, it seems to me that
a lot of times there's a "patsy" bull or two in the mix.
The short
go at the 2006 Tulsa Xtreme Bulls is a good example. Typically
in short rounds, the list of bulls is set, and then riders are drawn
to the bull list, while the list stays in order low to high. As
it happens, at Tulsa - 182 Hot Damn was drawn into the high position,
and the guy that won the long go was going to have him. So my guess
was that whoever won the long go was by far the favorite to win
the whole thing. Why? Hot Damn bucks off better bull riders less
than half the time. While he is typically marked pretty high, his
buckoff percentage is abysmally low - particularly against better
talent. What if 508 Lucky Strike had drawn into that position instead?
Then the guy that came in in the lead would have been given far
less chance of winning, since 508's buckoff % vs. better riders
is over 90%. In fact, having those two bulls together in a very
limited field (15 or less contestants) seems to make it more of
a drawing contest in my mind, and bull riding ought not to be a
drawing contest any more than is necessary. If we stick to that
maxim, then it makes sense to pay attention to such things when
setting stock for a very limited round of competition such as a
short go round. In a limited round, it should not be that hard to
come up with a set of bulls that are fairly even, and it would appear
that marking is not always the best measure to use. In a limited
round, what really counts the most is whether you make the whistle
or not - not whether some other guy is a point more than you. If
you buck off, everyone who rides will be 80+ points better. So having
bulls of close to equal difficulty AND good performance seems fair,
does it not?
In light of that, the idea behind PBS power rating is to measure
both a bull's marking AND his record against riders and top riders
all in one swoop. So bulls with a similar power rating ought to
be of similar caliber AND similar degree of difficulty. To use the
above example again, let's take a look at each bull's power rating
number from that short round- 11 outs from the top:
64.45, 68.41,
72.09, 88.86, 85.91, 78.55, 71.61,
75.16, 81.30, 82.34, 80.38,
The ones in bold got rode - which were the 1st, 2nd, and 4th lowest
power ratings in that round. About what you would expect. So, it
would appear that subbing out 4 bulls in that round for the 4 lowest
power ratings might have made for a more even pen. Likewise, subbing
8 lesser ones for the 8 highest would work too, but where else are
you going to put those bulls? They don't fit so well in the long
round either because there aren't enough of them.
PBS Power Rating is a weighted index of five 0 to 100 point measurables
that is ultimately shown as percentage of 100 points. 100 is the
maximum power rating.
The contributing factors are:
40% average marking (magnified to 100 point scale)
30% buckoff percentage vs. better riders
10% buckoff percentage
10% percentage of outs that are vs. better riders
10 % number of outs vs. better riders as a percentage of 25 (max
100%)
For the purposes of this rating better riders are defined as riders
who have at least 50 outs, and a riding percentage of 40% or more.
If a bull has been to the PBR at all, for average mark we use his
average marking from everywhere plus his average from the PBR divided
by 2 magnified to a 100 point scale. If he hasn't been to the PBR,
then we just use the average marking magnified.
In the PBS Power Rating RANKINGS, we limit it to bulls that have
10+ outs vs. better riders. So if you are just looking at bulls
at random, keep in mind bulls that have never faced better riders,
or not many of them... will have a low power rating. It may be that
after they have been around a while it could improve.
STATISTICALLY SIMILAR BULLS:
For each bull with 5+ outs in the database, you will find a link
mid page on the right side of his profile that says: Statistically
Similar Bulls. This will show you up to 150 bulls that have comparable
average marking, Power Rating, buckoff percentage vs. top riders,
and Winning Percentage. The range for winning percentage is a bit
wider than for the other three criteria. This tool is fairly accurate
at showing you bulls that would fit in a pen with the bull you query.
RIDERS:
Determining which riders are the best is just as complex. The PBR
uses a point system, and traditionally, the best bullriders are
considered to be those who have won the most money over a given
time period. We don't store money won for bull riders here, and
in truth, it's not the best indicator anyway.
We calculate a number of different stats for bull riders, including
riding percentage, which is a common stat - the number of qualified
rides for a given rider, expressed as a percentage of his total
attempts. It's kind of like batting average for baseball players.
It's affected by the quality and quantity of bulls faced. A somewhat
better way of looking at riding percentages is POINTS PER OUT -
which we sometimes refer to as PPO here. It's simply the sum of
a rider's points on qualified rided divided by the total number
of attempts - including zeros.
Number of 85+ scores and the number of qualified rides favors guys
who get on a lot, because they have more opportunity to record 85+
scores, but in general, the number of qualified rides plus the number
of 85+ scores is the largest determining factor as to whether a
guy qualifies for the NFR or not (i.e. wins more money).
Most rider stats are available on our rider
rankings page.
EVENTS:
My main question regarding individual events is: Which events are
featuring the best bulls? I have gone about this a couple of different
ways in the past, but the most accurate way to my thinking is to
measure the impact of the presence of the top 1500 bulls on any
one event. In the overall bull ranking table, the 1500th highest
marked bull is around 21 points. So it seems to me that the event
which features the largest percentage of outs by proven bulls that
average over 21 points or so is probably the event with the best
bulls. Other indicators, such as the average ranking of all the
bulls at an event, and the average power rating, are also useful
for comparison.
When comparing bull quality at different events, you have to look
at the career level quality of the bulls that were there. Looking
at the markings they recieve at the event can be misleading because
judges may have been marking high or low that day. All an event
producer can be expected to do is to bring the best bulls available
- they can't make them perform. So most of the event rankings here
correct themselves over time as the bull's individual careers play
out, and we look at them with the benefit of hindsight - years worth
of it.
Most event-related stats can be found on our event
rankings page. Also in the "at a glance" section under
each event.
GENERAL CONSIDERATIONS:
I haven't yet run across a situation in which MORE data would not
help. There is a woeful lack of data available to draw conclusions
from. We only have information dating back to January 2003, and
are working on various ways to get more. Some riders and bulls are
not treated fairly by our rankings, because their entire careers
are not included. The most obvious example of this is 61
Little Yellow Jacket, for whom we only have the tail end of
a brilliant career. He ranks third amongst all bulls with 10+ outs,
but would likely be first by a landslide if we had all his outs.
There are bulls that are overrated, and bulls that are underrated.
This is mostly due to the fact that judging is subjective. A bull
is more likely to rate incorrectly if he is seen by a very narrow
range of judges, or if he spends most his time in an inferior herd.
If a bull goes primarily to PBR events, his ranking is more likely
to be accurate, due to more diligent record keeping in the PBR.
PBR events almost always yield 95% or more usable data from the
judge's sheets. Oftentimes high-profile PRCA sanctioned events will
be missing a lot of bull data -
one or more judges fail to mark buckoff scores, times or whatever.
Sometimes it happens that the best-qualified judge at a particular
event will fail to mark the sheets, and so the data we do get comes
from the other guy.
Stats in general can be misleading if you are looking to validate
something you already believe. People have a tendency to see what
they want to see. To say that bull x was better than bull y on a
particular day is best left to the consensus of the qualified people
who witnessed it. Stats however, do provide us with a way of looking
at the REALLY BIG picture. They give us a perspective that is impossible
for us to grasp on our own. In looking at stats, you should approach
it like you were looking at the earth from outer space... if your
objective is to know what it feels like to contemplate the vastness
of the ocean, then the view from outer space is not the best perspective
- you should go stand on a beach. But if you wish to get some idea
of how much of the earth's surface is covered in water, the view
from space would be appropriate. Without it, a fellow that lives
in Kansas might tell you that "not much" of the earth
is covered in water, while a person that lives on Key West might
completely disagree with that notion.
Stats sometimes help to debunk popular myths - even the ones you
really want to believe. No sport makes more use of more statistics
than baseball, and the Godfather of modern baseball statistics -
Bill James - once summed it up like this:
"Virtually all sportswriters, I suppose, believe that
Jim Rice is an outstanding player. If you ask them how they know
this, they'll tell you that they just know; I've seen him play.
That's the difference in a nutshell between knowledge and bullshit;
knowledge is something that can be objectively demonstrated to
be true, and bullshit is something that you just 'know.' If someone
can actually demonstrate that Jim Rice is a great ballplayer,
I'd be most interested to see the evidence."
----------- The 1985 Bill James Baseball Abstract
Stats are not the only thing, but they are an important thing.
My approach is based on curiosity. I mentioned above that I start
with a question. Sometimes it is not the right question, and sometimes
it takes me a while to come up with the right question. This is
the nature of statistics. You have to know what questions can be
answered, and to what degree the answers might be accurate. They
do not take the place of human judgement or common sense, and they
shouldn't. What they should do is help us to see the big picture.
Slade Long - Probullstats.com - July 3, 2006
|
