Needless to say, there are lots of statistics. The endless discussions on Spencer are a case in point. No, he did not get a lot of sacks in his career. Yet, he got more this year and was named to the pro-bowl. More importantly, sacks are good, but not the ONLY measure or even the best measure of a LB.
One can read BTB and get a feel for how and when they are used. The best parts of many posts are actually in the discussion. Often someone will add a point and back it up by explaining why a given statistic is good or not for the discussion at hand.
This is the first of a series to discuss how to use statistics better without getting lost in formula. As Barbie says “Math is hard”, so actual calculation and math theory will be minimized.
A scale can be a great tool to figure out what you weigh, but does not provide information on how to LOSE WEIGHT. Like a scale, statistical analysis provides diagnostic tools. Used properly they can help a team figure out what they are doing wrong and show a direction to how to fix it.
Studies of quality control provide a methodology. Not surprisingly, the position for quality control is the guy who keeps the stats. Quality control guys are very good at using stats, and determining which to use based on what they measure. Second, they try to use the cause-effect relationships not just association or linkages.
The best diagnosis is to look at a complete list of symptoms and try to derive the cause. To cure an illness one has to identify and eliminate the underlying cause, not just treat the symptoms. Treating a symptom just hides the problem. One can take a cough drop for a sore throat caused by pneumonia. Your throat might not feel as bad for a short period of time, but you won’t get any better. You might actually get worse as it may allow you to rant more and cause more problems.
OTOH, take penicillin to kill the source and you can not only will get healthy, but eventually you will feel better. Tke kicker is that you might feel a lot worse until it starts to work.
Efficiency and absolute statistics
A single statistic by itself can’t tell a full story. For example, a firm generally uses a minimum of three stats – profit, ROI, and cash flow. Each is important, but one has to look at all three at the same time.
Everyone wants more profit, defined as the revenue coming into the firm minus cost, the money going out. Profit is an absolute number. The bigger the profit, the better it is. A 2 million dollar profit is greater than 1 million dollar profit.
Yet, consider the case of someone who earns a $10,000 profit. Not a big deal as compared to the big companies. Yet if you are a girl scout selling cook scout cookies, making that kind of profit gets you on the news. On the other hand Apple reported record profits this quarter and the price of its stock fell.
Profit is not the only issue. It also depends on what you invested to get that number. If it takes a big investment to get the profit, then some can’t get involved in the business in the first place.
Say you have an opportunity to invest in a startup firm A with the return of $2,000,000; if you invest $100,000. If you do not have that initial money, it doesn’t matter what the profit will be.
Say investment B only costs $ 50 yet gets the $1,000,000 return. Even though that is less profit, it would be better investment than investment A.
Unlike profit, return on investment [ROI] is a relative number as it measures how efficient one is in earning the profit. A return depends upon how much the initial investment was. All things being equal, one wants the higher ROI.
Yet things are not always equal. Say investment C gets 15% return and investment D gets 1,000%. If the return on C is a profit of $1,000 dollars that is still a bigger profit than if D only gets $50 total profit. D might be a kid selling his comic book collection. A third measure is cash flow. If you have more in the bank you survive for a while even if you are not profitable for a while.
Surprisingly many quickly growing firms are very profitable when they go out of business, they just don’t have the cash flow to continue. Maybe they have sold everything they have made, but their buyers have not paid for their products yet. When they run out of cash, the company cannot afford to buy new raw materials.
Still if you have a great ROI and a great profit, one might be able to go the bank and borrow money, even with interest, and continue.
Key point - All three together provide better information than any one by itself.
Back to football, often there are multiple things going on at the same time. The trick is to prioritize the problems and go through them systematically. Not for nothing, Garrett says work the process repeatedly.
Yards given up and points allowed certainly show if a Defense is doing well, but not as much as to why and how to fix it. Each of them has issues. Together, they are better than each individually as they have different issues.
The bend/don’t break defense can be very effective as teams tend to have execution errors over time. The longer that an offense has to go, the more yardage that they give up, but eventually they stop themselves. Unfortunately, Dallas offense showed that too well.
One fix is to continue longer drives. We were poor on 3rd downs and tried our best to avoid 3rd downs and often passed a lot on 1st down. Our running was mostly just to keep the Defense honest and the team used lots of finesse instead of power. That worked a great deal until we had 3rd down – then we needed a solid run when we needed it. Note how poorly we were in the RZ and to continue 3rd down.
We added a true FB in Vickers and Murray/Tanner has a different skill set than MBIII/Jones//Choice. The OL has gone through a huge change in the last two years.
Points given up will accrue if it is just bend/don’t break or if something else is the cause. If the offense turns over the ball at their own red zone, there is not much the D can do to prevent some points scoring – they can just try to limit the damage. The same occurs with the special teams, who can help or hinder a defense significantly.
Teams doing well on both indicate more than a team only good at one, and same for a team doing poorly.
Statistics are great for diagnostic reasons. There are lots of data collected, yet each one by itself is rather meaningless. The key is to sort them out and use them in conjunction with others to try to find cause and effect relationships.
In later articles in the series, we will look at how to improve the use of statistics.