Once again, I'm clearing out my inbox, two e-mails at a time. At this rate, it will be empty by 2026...
All right Ken, enough is enough. What is going on with Illinois' rating? I keep looking at them with every passing loss...I mean game...and they still remain a few notches above Purdue. I understand Illinois' preconference schedule was pretty tough. I guess what I'm wondering is does your formula put too much weight on strength of schedule? Because to be 10-12, I don't really care who you are, I don't think you should be rated 35 in any ranking system. They just look completely out of place in the context of the surrounding teams in your ratings. What gives?
This was written before Purdue's win over Illinois on Saturday, which dropped Illinois to 10-13 overall and below Purdue in the power ratings. However, the idea that Illinois is so highly ranked by my system (and I should note, by other predictive systems) offers a conundrum that most people around college hoops want to ignore. A team with a losing record can actually be better than a team with a winning record.
Any number used to describe a team's or player's performance need to be taken in context. A team's record requires the same burden. If Illinois had played Purdue's schedule, it's safe to say they'd have a better record. If the Illini had Purdue's lack of misfortune in close games, their record would be even better.
To its credit, the selection committee seems to understand this better than just about anybody. They've been roundly criticized for giving a 26-2 George Washington team an eight-seed as they did in 2006, or leaving a 25-3 Utah State out of the field entirely in 2004. We'll hear it a bunch of times over the closing week of the season, that if a certain team gets to X amount of wins, they have to get an at-large or even be seeded highly. Of course, they don't. The sheer number of wins or losses are irrelevant without know the circumstances surrounding those games. Should Illinois be considered worse than every power conference team with a winning record?
If you're looking for a system that orders teams based on how good their season has been, then my system is going to let you down. Illinois is not having a good season, but just because they've lost a bunch of games doesn't mean they're the equal of Michigan or Northwestern, either. Unlike those two teams, no Big Ten squad should take the Illini lightly.
I am not really a basketball fan, so please forgive me if my question seems idiotic. After reading your Memphis free-throw article, I thought perhaps you would be the best person to answer it.
In the final seconds of the Georgetown at West Virginia game on January 26, WVU held a 57-55 lead. GU had the ball and was trying to set for a final shot.
Once it got under 10 seconds remaining with GU still not set up, my admittedly inexperienced instinct was that WVU should foul before GU could begin to shoot. They did not, and GU took a three-point attempt that fell with six seconds remaining, resulting in a 58-57 GU victory.
My question is simply this: Which of the two strategies has a greater mathematical probability of success?
- Allow GU to shoot from the floor and possibly hit a (virtual) game-winning 3-pointer, or
- Intentionally foul GU before they start to shoot and once the clock got under 10 seconds
Before we take a look at this, let's get one thing straight: no coach is going to employ such a strategy. The vast majority of coaches refuse to give a foul on the last possession when they're up by three, which is a strategy that vastly improves the leading team's chances of victory. With that said, the strategy posed by Shane is worth discussing from a theoretical point of view.
One thing we shouldn't forget is that this question assumes that the players are able to carry out the strategy successfully, which in a free-flowing game like basketball is not a safe assumption. However, I'll make this assumption to simplify the analysis.
To start the calculations, the Mountaineers would have most likely fouled Jonathan Wallace or Jessie Sapp, the last two Hoya players to touch the ball on the possession. Wallace is an 85% shooter this season in just 25 attempts, 83% on his career. In any case, Wallace is a better shooter than Sapp, so let's use Wallace's figure to test the worst case from WVU's perspective. There are three possibilities related to Wallace's hypothetical trip to the line. (For those averse to intense math, you might want to turn away.)
Scenario A: Wallace makes both four shots. There's a 72% chance Wallace makes both free throws and ties the game. WVU would have had a few seconds for a final possession. I'll just guess and say they have a 40% of scoring. This may be a little high, but we saw their final possession, and there's a reasonable argument that they did score anyway. They only had six seconds then, but in the fouling scenario they'd have a little more time.
Secnario B: Wallace makes the first and misses the second. There's a 13% chance of this happening. Georgetown's chances of winning then depend on them getting an offensive rebound and making the follow-up. I'll generously give them a 25% chance of the offensive rebound and a 50% chance of converting.
Scenario C: Wallace misses the first. There's a 15% chance of this happening, in which case I'll assume that Wallace intentionally misses the second. The same percentages hold for scenario B from that point on, although in this case a conversion results in a tie. I'll throw in a 10% chance that the conversion results in an and-one, thus giving Georgetown the lead.
To sort out what these options mean in terms of winning, let's put in table form. The percentages listed are absolute. For example, there's a two percent chance of Scenario C occurring and Georgetown converting the second missed free throw for a tying basket.
Scenario A B C Sum
Occurring: 72% 13% 15% 100%
Georgetown tying: 72% 0% 2% 73%
Georgetown leading: 0% 2% <1% 2%
Then WVU winning: 29% 11% 13% 53%
Then WVU tying: 43% 0% 2% 45%
From the WVU tying scenario, it seems reasonable to assume that overtime is a 50/50 proposition in this case. So I'll put half of that 45% into both the WVU and Georgetown win chances. This leaves us with roughly a 75% that WVU wins by employing this strategy.
Then the question becomes whether the Mountaineers would have a better chance of winning by playing the final possession straight up. I'm not going to bore you with the details of that calculation, but it's a close call. The two most variable factors in this strategy are the free throw shooting of the opposing player and the chances of winning in overtime. If UTEP's Tony Barbee had encountered this scenario in last Saturday's surprisingly close finish against Memphis, it would have made a lot of sense for UTEP to foul. He would not have done so, of course. Given how far this type of strategy is from mainstream thinking in college hoops, it will only ever be employed in video games. Regardless, there are cases where it would come in handy in real life.
Ken Pomeroy is an author of Basketball Prospectus.
You can contact Ken by clicking here or click here to see Ken's other articles.