Pay No Attention to the Formula Behind the Curtain

Hello Internet! I realize that this post going up on Thursday night means that I'm posting Thursday night then Friday morning. I had a crazy week. Also, they're two totally different subjects, so they get different posts. You know, for posterity, or something. Ok: Feast of F1 Data, Part I:

Race Finish	Driver	Starting Position	Laps Led	Laps Completed	Old Formula	Ranking
1	Alonso	3	31	56	82.22	1
2	Raikkonen	2	0	56	47.73	4
3	Hamilton	1	4	56	44.48	5
4	Vettel	9	6	56	59.90	2
5	Button	8	6	56	53.08	3
6	Massa	5	1	56	37.26	7
7	Ricciardo	7	0	56	36.36	9
8	di Resta	11	0	56	40.91	6
9	Grosjean	6	0	56	25.00	13
10	Hulkenberg	10	8	56	36.69	8
11	Perez	12	0	56	29.55	11
12	Vergne	15	0	56	31.82	10
13	Bottas	16	0	56	29.55	12
14	Maldonado	14	0	56	20.45	16
15	Bianchi	18	0	55	24.88	14
16	Pic	20	0	55	24.84	15
17	Chilton	19	0	55	18.10	17
18	van der Garde	21	0	55	18.06	18
19	Rosberg	4	0	21	-3.69	22
20	Webber	22	0	15	8.04	19
21	Sutil	13	0	5	2.92	20
22	Guitierrez	17	0	4	1.46	21

So, this is the Old Formula, a true Scoring Indy original. Things that are both spectacular and wonderful about it are as follows:

1) The re-shuffle of the Top 5 is perfect. Vettel's tire (sorry, F1, tyre) strategy worked out as well as could be expected, and I'd say that this was Button's best drive of the year.

2) The rule of thumb that 35 and up is a good drive holds, with Race Score ranks 1-9 scoring above the 35 point threshold.

I suppose that this boils down to, it basically gets the job done and I know how to read it.

Here are the things that bug me:

1) Rosberg was doing really well when his suspension went. Not his fault. I don't want to penalize people for stuff that isn't their fault. And yet, here's Nico Rosberg with the only negative score.

2) Massa seems high. Now, this is splitting hairs, but the point of scoring these races is to determine who did better than whom. Just, instinctively, I want to say that Hulkenberg and Ricciardo did better.

So, let us now take a look at possible replacements and see how our added data from lap charts and running orders can help to clear up our annoyances, while still maintaining the things that we like about the old formula. Here come the Feast of F1 Data, Part II:

Race Finish Driver Starting Position Laps Led Laps Completed Formula A Formula B Formula C 1 Alonso 3 31 56 86.14 95.47 95.47 2 Raikkonen 2 0 56 61.39 73.84 73.84 3 Hamilton 1 4 56 58.81 76.88 74.21 4 Vettel 9 6 56 66.38 72.60 68.60 5 Button 8 6 56 63.04 70.30 64.97 6 Massa 5 1 56 44.36 64.91 58.24 7 Ricciardo 7 0 56 44.45 64.92 62.25 8 di Resta 11 0 56 50.62 57.48 53.48 9 Grosjean 6 0 56 32.42 56.09 49.42 10 Hulkenberg 10 8 56 47.21 61.76 49.76 11 Perez 12 0 56 40.86 52.86 42.20 12 Vergne 15 0 56 45.96 51.49 43.49 13 Bottas 16 0 56 35.30 39.69 34.36 14 Maldonado 14 0 56 23.58 36.69 31.35 15 Bianchi 18 0 55 29.52 28.81 27.51 16 Pic 20 0 55 25.99 28.23 26.92 17 Chilton 19 0 55 19.66 24.90 22.28 18 van der Garde 21 0 55 18.65 18.65 17.34 19 Rosberg 4 0 21 4.95 19.41 12.41 20 Webber 22 0 15 10.15 12.02 8.45 21 Sutil 13 0 5 3.14 5.08 3.77 22 Guitierrez 17 0 4 2.29 2.83 1.78

So, here are brief descriptions of the three formulae:

- Formula A looks a lot like our own formula. What it adds is that it takes into account average running position as an equal factor to race finish and gaining positions.

- Formula B goes even further. Instead of working with positions gained against starting position, this Formula looks at positions gained from the lowest that driver ever was during the race to the highest position that the driver ever had officially scored at the end of a lap.

- Formula C tries to be a middle road. It accounts for positions gained from the lowest running spot to finishing position.

How do they address the concerns above?

- Shakes up the Top 5 into the order: Alonso, Vettel, Button, Raikkonen, Hamilton: Formula A: YES / Formula B: NO / Formula C: NO

- Keep the rule of thumb that 35 points is a good drive: Formulae A, B, & C: NO (Note that this isn't technically a problem, we'll just have to re-calibrate what "good" is.)

- Does something to keep Rosberg out of the cellar: Formulae A, B, & C: YES (I think that Formulae B and C do a little TOO much to help Rosberg. He only ran 21 laps, after all, and Formula B puts him AHEAD of van der Garde. Aside, van der Garde's first name is Giedo. That's pretty close to Guido. He's my new favorite. Also, the livery on those Caterhams looks cool.)

- Puts Hulkenberg and Ricciardo ahead of Massa: Formula A: YES / Formula B: NO / Formula C: NO (Aside, both B and C put Ricciardo above Massa, but Hulkenberg stays behind.)

So where does this leave us? Well, right now I like Formula A best for the new Post-Indy Scoring Indy formula. That's not in pen yet. I have to finish doing all the data entry for the IndyCar races up to this point (as well as checking my results against the yet-to-be-contested Long Beach and Sao Paulo races). We'll know for sure by Indy.

Long Beach preview goes up tomorrow! Stay tuned for the Scoring Indy Twitter predictions after qualifying!

Later!

-- Guido

Behind the Curtain on the New Formula

Greetings Internet! I missed you. Did you miss me? I hope so. Anyway, here's the Friday update.

With no immanent IndyCar date (Long Beach is another week away), I have some space to fill. You may remember from my Barber recap that with access to lap charts, I can make some changes to my scoring formula. Here's a brief look at what ScoringIndy is cooking up:

The factors that are considered under the current formula are this:

1) Race Finish
2) Positions Gained/Lost
3) Laps Led
4) Laps Completed
5) Qualifying v. Starting Position

Under the current formula, Race Finish accounts for roughly half the scoring, whereas the other factors are cooked together to create some kind of "In-Race Performance" term that accounts for the other approximate half.

Now that we can add average running position to the mix, we have three terms to our mathematical jamboree:

1) Race Finish
2) Average Running Position
3) In-race Performance

So, the question that we ask ourselves is this: what is the relative weight of each of these terms? How much credit does each receive in the mathematical pressure-cooker that is our formula? We also need to ask ourselves if there is anything that can be done to enhance the In-Race Performance term using the smorgasbord of lap chart data.

Now, I'm not sure what the answers are to these questions. And, that is one of the reasons that I'm looking forward to Formula 1's visit to Shanghai this weekend. The smaller number of cars and laps will make data entry easier for me, and that will allow for more time spent tweaking numbers. At some point next week, I will be posting various scores for the weekend's F1 race using a variety of different formulae, and this will let us all take a look at what we may want to do for a Race Score Formula going forward. I'll also (in the privacy of the ScoringIndy lab) be adding the lap chart data to my IndyCar spreadsheets, and tinkering with formulae there.

Remember, a new formula will be in place for Indy.

I know this week was short. Don't worry, we'll have F1 New Formula Goodness soon!

-- Guido

Looking Back on Bama

Greetings Internet! I hope you all enjoyed yesterday's race-itude as much as I did. Compelling action on track, and the grass is actually green in Alabama. (It gives me a little hope for Wisconsin... Maybe by the time Milwaukee comes around...)

Anyways, I know what you're here for. Here's the gala of data:

Race Finish	Driver	Starting Position	Laps Led	Laps Completed	Race Score
1	Hunter-Reay	1	53	90	79.44
2	Dixon	4	0	90	51.92
3	Castroneves	6	25	90	65.81
4	Kimball	5	3	90	47.82
5	Power	2	9	90	41.54
6	Pagenaud	13	0	90	53.85
7	Andretti	7	0	90	38.46
8	Wilson	8	0	90	36.54
9	Newgarten	22	0	90	59.62
10	Vautier	3	0	90	19.23
11	Tagliani	15	0	90	38.46
12	Viso	16	0	90	36.54
13	Kanaan	19	0	90	38.46
14	Sato	12	0	90	21.15
15	Servia	18	0	90	28.85
16	Bordais	23	0	90	34.62
17	Hildebrand	24	0	90	32.69
18	de Silvestro	14	0	90	9.62
19	Allmendinger	10	0	90	-1.92
20	Saavadra	9	0	90	-7.69
21	Rahal	21	0	90	11.54
22	Carpenter	26	0	89	17.22
23	Jakes	11	0	86	-14.36
24	Beatriz	25	0	65	7.16
25	Franchitti	17	0	42	-3.33
26	Hinchcliffe	25	0	3	1.86

Quick Observations:

- Hunter-Reay (79.44); Castroneves (65.81); and Newgarten (59.62) on the Race Score Podium.

- The only interesting one of those three is Josef Newgarten, who charges from P22 to P9 to grab the third spot. I'd say this is sort of the expectation at Barber, not a ton of passing in the front, so someone working 13 spots up the grid like Newgarten did in his #67 Honda is worth recognizing.

- Lots of Race Scores in the 30s and 40s. You almost expect that with Barber Motorsports Park. As I alluded to earlier, with such a technical course, it's hard to gain or lose lots of positions, which is what makes Newgarten's run amazing.


Twitter Predictions:

- #HIGPA Prediction 1) Kimball is this race's Simona: good car, but won't improve position much, and gets a lower Race Score because of it.

- VERDICT: Mostly on. Kimball starts P5; finishes P4; has a Race Score of 47.82 (outside the top 5). The only difference between his run, and Simona's run last week is that he moved up a spot, whereas she moved down two. And, that kept him in the Race Score World Championship points.

- #HIGPA Prediction 2) Pagenaud and Hinch get big Race Scores. Those teams know what to do and they're farther back than they should be.

- VERDICT: Halfway there. Pagenaud makes me feel smart. Hinch makes me cry. I will say, for the record, that was not Hinchcliffe's fault.

- #HIGPA Prediction 2a) Add Franchitti in pencil to prediction 2. In pencil because I don't trust him to not to ruin his own day again.

- VERDICT: See Hinchcliffe. Franchitti was having a very good drive, but something broke on the #10 Honda, and that ended his day. Thank goodness for Pagenaud. Again, this was a good prediction, and Franchitti didn't ruin his own day, so I feel good about it.

- #HIGPA Prediction 3) A Honda wins, probably Dixon.

- VERDICT: Nope. The Hondas were good, with 6 of the top 10; but I should have known better than to pick Scott Dixon at Barber. He finishes second for the fourth year running. Also, he's a foolproof trivia answer. If anyone asks you "Who finished second at the IndyCar race at Barber in..." you can cut them off before they say the year. The answer is Scott Dixon. Also, Dixon gets quote of the week for suggesting that he go buy a bridesmaid's dress after four straight years in second.

- #HIGPA Prediction 4) Castroneves leaves Alabama with the lead in the Race Score World Championship. (Mostly because of Turbo the snail)

- VERDICT: Nailed it. Fear the Snail.


R.S.W.C. Update

Here are the updated standings in the Race Score World Championship:

Driver	RSWC Points	Points Back
Castroneves	43	0
Dixon	28	-15
Hunter-Reay	25	-18
Newgarten	15	-28
Viso	15	-28
Pagenaud	12	-31
Andretti	12	-31
Kanaan	11	-32
Kimball	8	-35
Power	6	-37
Tagliani	6	-37
Bordais	6	-37
Wilson	2	-41
Carpenter	1	-42

Observations:

- Interesting that the Top 3 in the RSWC are the Top 3 in IndyCar series points (in the same order, no less).

- As I said before, fear the snail.

-  Drivers in the top half of the IndyCar standings that don't have any points in the RSWC: Simona de Silvestro and Takuma Sato. They won't stay off this list for long.

- Ed Carpenter will finish in the Top 10 for the RSWC. I'm sticking by it. But, I don't know that E.J. Viso will.


Announcement:

In my pokings around on the little corner of the internet devoted to open-wheel racing, I've discovered that MoreFrontWing.com (shameless and unsolicited plug for a great IndyCar site) has lap charts for IndyCar events. This will allow me to incorporate some other things into my Race Score formula. Most notably, it lets me play with average running position.

This blessing, however, is also a curse. You see, computing average running position from just the data that I have means that I have to enter lap-by-lap data into a spreadsheet. I'm willing to do this, but it's time consuming. This means that I'm not going to have a new formula done immediately, due to the requirements of needing to test it to make sure it does what I want it to do: generate accurate and thought-provoking insights about drivers' performance in races.

Therefore, the announcement is this: starting with the 2013 Indy 500, Scoring Indy will be using a formula that accounts for average running position (as well as the things, for which the current formula accounts). I need all the time up to Indy to get the formula right.

Stay tuned to the blog and @ScoringIndy on Twitter for updates on the new formula. We might try out some different stuff using the F1 results from China. That could be a fun project for next week...

That's all I've got for now. I won't be gone long.

-- Guido

A Trip Back to the Barber Shop

Welcome back, internet! (Or, should I be welcoming myself back out of my Holy Week cloister?) In any event, I returneth, and just in time for Barber, too. So, to get us all prepared for the Elevation-Change Circus that is Round Two of the IndyCar schedule, here is last year's hurley-burley of data:

Race Finish	Driver	Starting Position	Laps Led	Laps Completed	Race Score
1	Power	9	22	90	77.61
2	Dixon	3	38	90	71.11
3	Castroneves	1	28	90	57.86
4	Rahal	8	1	90	52.48
5	Pagenaud	10	0	90	51.92
6	Hinchcliffe	2	1	90	33.25
7	Conway	4	0	90	32.69
8	Barrichello	14	0	90	48.08
9	Bourdais	17	0	90	50.00
10	Franchitti	18	0	90	48.08
11	Andretti	13	0	90	34.62
12	Hunter-Reay	11	0	90	26.92
13	Servia	26	0	90	51.92
14	Briscoe	12	0	90	21.15
15	Hildebrand	5	0	90	3.85
16	Jakes	20	0	90	28.85
17	Newgarten	15	0	90	15.38
18	Viso	7	0	89	-3.61
19	Wilson	19	0	89	15.38
20	de Silvestro	21	0	89	15.36
21	Kanaan	6	0	89	-16.99
22	Carpenter	23	0	88	11.50
23	Legge	24	0	85	9.51
24	Sato	16	0	52	-3.12
25	Kimball	22	0	45	0.96
26	Tagliani	25	0	0	1.92

So, here are my thoughts:

- It seems that you can pass at Barber now. (Compare last year with 3 drivers leading 20+ laps to 2011, when Will Power lead every lap from the pole: a 100 point Race Score, and you quickly see my point.) So, we could see some people putting up big numbers by coming up through the field. See Viso and Dixon in St. Pete, and Power last year here at Barber.

- Is it possible for Will Power to run for Mayor of Leeds, Alabama? He has two of the three poles and two of the three wins. Expect him to have a car to be right up there again this weekend. I'd go so far as to say that he should win, provided he doesn't get run over again. Don't text and drive! (Yes, I know I'm going way out on a limb here...)

- Sam Schmidt's car looked good last year at Barber with Pagenaud. Vautier looked very good in his first IndyCar Series race. At least one of them ought to be quite competitive this weekend.

- KV also looked good last year in Alabama and two weeks ago in Florida. TK and Simona are in a good position to make real noise at the Alabama GP.

- I'm excited to spend the weekend saying the word "Allmendinger."

- Finally, stay tuned to @ScoringIndy on Twitter. After qualifying, I'll be posting some race predictions, as well as predictions specific to the Race Score World Championship.

That's all I've got for today, I'll be tweeting this weekend, and blogging on Sunday night or Monday morning. Let's go racing, Alabama!

-- Guido