Tourney Projections: A few new features

I added the following features with regards to NCAA tournament bracket projections:

1. ON the main bracket page (www.rpiforecast.com/bracket.html), I added more stats: Current and Expected RPI, Expected OOC RPI, Current and Expected Records, and Current and Expected Conference Records

2. The daily projected tournament seed can now be found on the individual team pages (for example, see Duke: www.rpiforecast.com/teams/Duke.html)

3. I've added a graph on the individual team pages below the other two graphs which tracks the history of projected seeds for each team. You will be hard pressed to find others willing to archive the history of their projections.

Next on the to-do list is BracketBusters and Conference Tournaments.

New Updated Bracket predictions (Daily!)

Based on the new updated "Dance Card" formulas from Jay Coleman, I have started making daily predictions of the field of 68 along with the seeds. A great thing about the methodology is that the factors that determine the at-large bids and the seeds are similar but there are subtle differences. The new model only missed one team from the field and did as well as pretty much anybody at picking the actual seeds. I will be keeping track of my predictions as well as those of the other major bracketologists through the course of the season. After the season is over, I'll present results. If the past is any indicator, my predictions should do especially well early in the season compared to the "human predictions".

Find the predictions here:

http://www.rpiforecast.com/bracket.html

Live-RPI.com - Minute by Minute RPI Updates - Free!

You may have noticed a brand new feature this season: live-rpi.com . This new website is for those who don't want to wait until the next morning (or even 5 minutes) to see what the current RPI is. As games finish, the RPI is automatically updated and the page should refresh. If it doesn't, simply click on the Live-RPI.com banner and it should refresh. From about 6 am EST to around 1pm, there are no updates unless games are being played. There is also a section on conference RPI which is updated at the same time.

As always, this is free!

Brackets!

How'd we do in projecting the final bracket? Not quite as well as previous years. This is disappointing because in past years, the tournament projections have been more accurate than the vast majority of the other prognosticators - especially from November through February. The bracket forecasts are based on Jay Coleman's OLD "Dance Card Formula". This formula was estimated quite a few years ago to model the behavior of the selection committee. Inevitably, the selection committe changes its behavior over time, so maybe this is not a huge surprise.

The good news is that Jay Coleman has just published a new version of his model. In fact, this new version predicted all but one of the at large bids - and the team missed (Florida versus Miss. St.) were ranked right next to each other and had very similar scores. For next season, we will use the improved updated formula and will also start projecting seeds instead of simply ranking them in order of probability of an at-large bid. Jay Coleman and his co-authors also have a new model that predicts seeds which will be used on this site.

Most pundits project brackets using a "what if the season ended today" framework. The approach taken on this site is to project what the field will look like on Selection Sunday. The advantage of combining the simulations on RPIforecast.com and the Dance Card formula is twofold: (1) The actual forecasts of RPI and records which feed into the formula are much more accurate than using the current RPI and record and (2) The Dance Card Formula (especially the new version) is objective and unbiased and does a remarkable job at predicting the field.

Home Court Advantage

One of the most important aspects of basketball - Home Court Advantage - is also one of the most difficult to measure. There are a few ways to do this. The most common is basically to estimate how many points playing on your home court is worth. Having teams play home-and-away in one season makes this possible. You could look at other variables such as Field Goal Percentage, but those may be affected by many factors besides the venue.

An alternative that came to mind is free throw shooting percentage. The idea is that because you can't really defend the free throw, there are only a few things should affect it: 1) who is shooting, 2) luck, and 3) the fans making noise and other distractions. To take care of number 1, we need to make sure to look not at the total free throw percentage of the road team, but instead look at who is shooting and how they do. That leaves us with luck and actual distractions from fans.

With this in mind, I wanted to figure out for each road team, what would you expect the free throw percentage to be based on who is shooting, and in doing so, you can measure the effect of the venue itself on performance. For those that have some knowledge of stats, I ran the following regression:

FTP_ik = a_i + venue_k + e_ik

where
FTP_ij is the free throw percentage of road player i in venue k
a_i is the unconditional average road free throw percentage of player i
venue_k is the effect of the home team on player i's free throw percentage
e_ik is the error term

venue_k is the thing I'm trying to measure. One more thing, I'm using Free throw attempts as a weight in the regression to give more weight in the estimation to those that shoot more often.

Below is what I came up with. The FT-Effect number is the (venue_k) estimate from above. For UME (Maine), is is -11.2%. That visiting players shot 11.2% lower Free Throw percentage against Maine than they would have in the average road venue. So, Maine it is! See where your team is:


Rank Home Team FT-Effect
1 UME -11.2%
2 DEST -10.6%
3 ORU -10.0%
4 NORF -9.9%
5 SVST -9.6%
6 MICH -9.2%
7 UVA -9.0%
8 SJSU -8.4%
9 FDU -8.4%
10 NEB -8.0%
11 GONZ -7.6%
12 BYU -7.5%
13 SUU -7.4%
14 USC -7.2%
15 CINC -6.9%
16 IONA -6.9%
17 DUQ -6.9%
18 MASS -6.6%
19 SYR -6.4%
20 LAF -6.1%
21 EMU -6.0%
22 TCU -6.0%
23 FGCU -5.5%
24 SOU -5.5%
25 GT -5.5%
26 COST -5.4%
27 UCR -5.4%
28 BETH -5.4%
29 RAD -5.3%
30 TROY -5.1%
31 WSSU -4.9%
32 PEPP -4.6%
33 UTA -4.5%
34 BOIS -4.4%
35 VAND -4.4%
36 USF -4.4%
37 LBSU -4.4%
38 GMU -4.4%
39 WVU -4.1%
40 VCU -4.1%
41 BING -3.9%
42 UNF -3.9%
43 IOWA -3.8%
44 UD -3.8%
45 CITA -3.7%
46 PITT -3.7%
47 MORE -3.7%
48 SIE -3.7%
49 PORT -3.6%
50 ELON -3.6%
51 WICH -3.6%
52 SPC -3.6%
53 NOVA -3.5%
54 WYO -3.4%
55 UNC -3.4%
56 SJU -3.4%
57 SC -3.4%
58 WASH -3.4%
59 ETSU -3.4%
60 ARMY -3.3%
61 NWSU -3.3%
62 AFA -3.2%
63 KENN -3.2%
64 TEX -3.2%
65 ID -3.2%
66 UCF -3.2%
67 MTSU -3.1%
68 TOWS -3.0%
69 ODU -3.0%
70 LAM -3.0%
71 UMES -3.0%
72 UNCW -2.9%
73 PSU -2.9%
74 LIP -2.9%
75 USU -2.9%
76 FSU -2.8%
77 SIUE -2.8%
78 SCL -2.8%
79 UNI -2.8%
80 SDAK -2.8%
81 BUFF -2.7%
82 AMER -2.7%
83 WKU -2.6%
84 OHST -2.6%
85 BSU -2.6%
86 CHAT -2.6%
87 NIAG -2.4%
88 VALP -2.4%
89 CLEM -2.4%
90 VMI -2.2%
91 USM -2.2%
92 CCU -2.2%
93 MON -2.1%
94 USD -2.0%
95 IND -2.0%
96 CAL -2.0%
97 LONG -2.0%
98 SCST -2.0%
99 WOFF -2.0%
100 SCU -1.9%
101 LTU -1.9%
102 QUIN -1.9%
103 MORG -1.8%
104 OU -1.8%
105 YSU -1.8%
106 GW -1.7%
107 UCSB -1.7%
108 SLU -1.6%
109 UCD -1.6%
110 UWM -1.6%
111 NMSU -1.6%
112 MOST -1.5%
113 PVAM -1.5%
114 NM -1.5%
115 SEMO -1.5%
116 UTAH -1.4%
117 IUPUI -1.4%
118 XU -1.4%
119 NIU -1.4%
120 INST -1.4%
121 STET -1.4%
122 NW -1.4%
123 POLY -1.3%
124 PRIN -1.2%
125 CAMP -1.2%
126 FURM -1.2%
127 BONA -1.1%
128 PUR -1.1%
129 ARIZ -1.1%
130 BC -1.1%
131 SAMF -1.0%
132 HAMP -0.9%
133 FAU -0.9%
134 FORD -0.9%
135 DUKE -0.9%
136 ASU -0.9%
137 NCAT -0.9%
138 FRES -0.8%
139 UIC -0.8%
140 UNCG -0.8%
141 LOYO -0.8%
142 SIU -0.7%
143 NAU -0.7%
144 NICH -0.7%
145 JSU -0.7%
146 SMC -0.7%
147 COFC -0.6%
148 ORE -0.5%
149 WRIG -0.5%
150 MSM -0.5%
151 PROV -0.5%
152 CANI -0.5%
153 MIO -0.5%
154 UMT -0.5%
155 UCLA -0.4%
156 HART -0.4%
157 EVAN -0.4%
158 MINN -0.4%
159 SDST -0.4%
160 KENT -0.3%
161 ALA -0.3%
162 MAN -0.3%
163 TUL -0.3%
164 COLO -0.3%
165 HAW -0.2%
166 LIUB -0.2%
167 JAX -0.2%
168 RID -0.1%
169 NCC -0.1%
170 WAG -0.1%
171 NCST -0.1%
172 ALC 0.0%
173 SDSU 0.0%
174 CORN 0.0%
175 UTEP 0.0%
176 UNH 0.0%
177 BELM 0.1%
178 MURR 0.1%
179 URI 0.2%
180 CONN 0.2%
181 SB 0.2%
182 NOCO 0.2%
183 WAKE 0.2%
184 FIU 0.2%
185 WISC 0.2%
186 TNT 0.2%
187 LEH 0.3%
188 FAMU 0.4%
189 DEP 0.4%
190 UTPA 0.4%
191 LSU 0.4%
192 WIU 0.4%
193 CHSO 0.4%
194 NAVY 0.4%
195 LIB 0.5%
196 SAMH 0.5%
197 WEB 0.5%
198 NJIT 0.6%
199 KU 0.6%
200 TXST 0.6%
201 MERC 0.7%
202 HOU 0.7%
203 KSU 0.7%
204 MTST 0.7%
205 ILST 0.7%
206 SFNY 0.8%
207 BAY 0.8%
208 TXSO 0.9%
209 BUTL 0.9%
210 DEN 1.0%
211 UVU 1.0%
212 UMKC 1.0%
213 ALAM 1.1%
214 HBU 1.1%
215 UMD 1.1%
216 UCA 1.1%
217 WSU 1.2%
218 UDM 1.2%
219 GWB 1.2%
220 OSU 1.2%
221 CCSU 1.2%
222 CLST 1.3%
223 TAM 1.3%
224 COLG 1.3%
225 RUTG 1.4%
226 PENN 1.4%
227 NTEX 1.4%
228 MEM 1.4%
229 TTU 1.4%
230 WMU 1.4%
231 STJ 1.5%
232 GU 1.5%
233 ILL 1.5%
234 AUB 1.5%
235 SMU 1.5%
236 UVM 1.6%
237 UAPB 1.7%
238 UWGB 1.7%
239 CST 1.8%
240 MAR 1.8%
241 COLU 1.8%
242 DART 1.9%
243 CENT 1.9%
244 FLA 1.9%
245 MARQ 1.9%
246 VT 1.9%
247 COPP 1.9%
248 IDST 2.0%
249 DREX 2.0%
250 APST 2.0%
251 UNCA 2.0%
252 ND 2.1%
253 EWU 2.1%
254 GS 2.2%
255 EIU 2.2%
256 CWM 2.2%
257 ECU 2.3%
258 CMU 2.3%
259 UALR 2.4%
260 UNLV 2.4%
261 GRAM 2.4%
262 OAK 2.5%
263 SF 2.5%
264 ARST 2.5%
265 NEV 2.6%
266 UTM 2.6%
267 UL 2.7%
268 MCN 2.7%
269 BUCK 2.7%
270 ULL 2.8%
271 APSU 2.8%
272 ALB 2.8%
273 HOFS 2.9%
274 ISU 2.9%
275 TOL 3.0%
276 UGA 3.1%
277 RICH 3.1%
278 NDST 3.2%
279 UTSA 3.2%
280 LAS 3.2%
281 STAN 3.2%
282 MVSU 3.2%
283 MSU 3.2%
284 IPFW 3.3%
285 GSU 3.3%
286 MIZZ 3.4%
287 MIA 3.4%
288 TAMC 3.6%
289 ULM 3.7%
290 DRKE 3.7%
291 BRY 3.7%
292 HP 3.8%
293 UNO 3.8%
294 PAC 3.9%
295 CREI 3.9%
296 MSST 3.9%
297 NU 3.9%
298 HC 4.0%
299 NDAK 4.0%
300 ORSU 4.0%
301 TU 4.1%
302 CSF 4.1%
303 PRES 4.1%
304 MISS 4.2%
305 RICE 4.4%
306 UAB 4.4%
307 FAIR 4.4%
308 HOW 4.5%
309 BRWN 4.5%
310 SH 4.5%
311 BGSU 4.6%
312 YALE 4.6%
313 TENN 4.8%
314 AKR 4.8%
315 JMU 4.9%
316 WINT 5.0%
317 JXST 5.2%
318 SFA 5.2%
319 SHU 5.3%
320 MRSH 5.3%
321 DEL 5.4%
322 OHIO 5.5%
323 UK 5.5%
324 LMU 5.5%
325 UCI 5.6%
326 SEA 5.6%
327 TSU 5.7%
328 DAV 6.0%
329 SFPA 6.1%
330 USA 6.3%
331 LMD 6.4%
332 CHAR 6.4%
333 RMU 6.6%
334 ARK 6.6%
335 EKU 6.7%
336 CSN 6.7%
337 SELU 6.7%
338 SAC 6.8%
339 ALST 7.7%
340 TULS 7.9%
341 BU 8.1%
342 WCU 8.1%
343 UMBC 8.2%
344 CSB 8.3%
345 POST 8.6%
346 BRAD 9.6%
347 HARV 10.3%



These numbers are all relative to the effect of the average venue, not relative to performance at home. I didn't include free throw percentage of any players at home in the estimation.

Here are the same estimates using equal weights (ignoring how many free throw attempts a player had):

Rank Venue FT-Effect
1 SVST -14.3%
2 UME -12.0%
3 NORF -11.8%
4 UNF -11.0%
5 MASS -10.9%
6 FDU -10.8%
7 BYU -10.2%
8 SUU -10.0%
9 MICH -10.0%
10 NEB -9.9%
11 GT -9.8%
12 CINC -9.7%
13 UVA -9.6%
14 AFA -9.0%
15 COST -8.7%
16 UCR -8.6%
17 PORT -8.5%
18 WYO -8.1%
19 LAF -7.9%
20 MORE -7.0%
21 SYR -7.0%
22 SOU -7.0%
23 IONA -7.0%
24 USC -6.8%
25 UIC -6.8%
26 DEST -6.8%
27 UVU -6.6%
28 IOWA -6.6%
29 PRIN -6.3%
30 BETH -6.2%
31 EMU -6.0%
32 SIUE -6.0%
33 GONZ -5.9%
34 SPC -5.9%
35 VAND -5.9%
36 PEPP -5.8%
37 UMT -5.8%
38 QUIN -5.7%
39 WICH -5.7%
40 WASH -5.5%
41 ODU -5.5%
42 SMC -5.3%
43 ARMY -5.3%
44 SCL -5.2%
45 ID -5.2%
46 BOIS -5.1%
47 TROY -4.9%
48 DUKE -4.9%
49 ASU -4.9%
50 PSU -4.9%
51 TCU -4.8%
52 UNI -4.8%
53 FGCU -4.8%
54 SJU -4.7%
55 ORU -4.7%
56 ALA -4.7%
57 WVU -4.7%
58 UD -4.7%
59 MSM -4.5%
60 VALP -4.5%
61 ETSU -4.4%
62 BSU -4.4%
63 GMU -4.3%
64 SDSU -4.3%
65 IUPUI -4.1%
66 UNC -4.1%
67 NIU -4.1%
68 SCST -4.0%
69 RID -4.0%
70 SDAK -4.0%
71 SJSU -3.9%
72 HART -3.9%
73 HBU -3.8%
74 KENN -3.7%
75 ELON -3.7%
76 UTA -3.6%
77 ORE -3.6%
78 NOCO -3.4%
79 NOVA -3.3%
80 VMI -3.3%
81 XU -3.3%
82 FSU -3.2%
83 UCSB -3.2%
84 BC -3.1%
85 OAK -3.1%
86 GW -3.1%
87 NJIT -3.1%
88 YSU -3.0%
89 LBSU -3.0%
90 SIU -2.9%
91 UCF -2.9%
92 NCAT -2.9%
93 RAD -2.8%
94 BING -2.8%
95 SEMO -2.8%
96 SDST -2.8%
97 VCU -2.8%
98 MTSU -2.8%
99 AMER -2.8%
100 MTST -2.7%
101 DUQ -2.6%
102 WKU -2.6%
103 LAM -2.6%
104 STJ -2.6%
105 SLU -2.4%
106 SB -2.3%
107 NIAG -2.3%
108 LONG -2.3%
109 NMSU -2.2%
110 CHAT -2.2%
111 UNCG -2.2%
112 MORG -2.2%
113 KENT -2.2%
114 EVAN -2.2%
115 CHSO -2.1%
116 CAL -2.1%
117 TEX -2.1%
118 USF -2.1%
119 MON -2.0%
120 LIB -2.0%
121 UTPA -2.0%
122 URI -2.0%
123 CAMP -1.9%
124 NAU -1.9%
125 BUFF -1.9%
126 CANI -1.8%
127 UMES -1.8%
128 NAVY -1.8%
129 NM -1.7%
130 PITT -1.7%
131 CLEM -1.7%
132 LAS -1.6%
133 WOFF -1.6%
134 STET -1.6%
135 UVM -1.6%
136 DEN -1.5%
137 MARQ -1.5%
138 UCD -1.5%
139 SIE -1.4%
140 INST -1.3%
141 WSSU -1.3%
142 UWM -1.3%
143 LEH -1.2%
144 WAKE -1.1%
145 MOST -1.1%
146 USU -1.1%
147 MERC -1.1%
148 FAMU -1.1%
149 SC -1.0%
150 CITA -1.0%
151 FAIR -1.0%
152 LIP -1.0%
153 CST -0.9%
154 VT -0.9%
155 COFC -0.9%
156 UWGB -0.9%
157 UDM -0.7%
158 NICH -0.7%
159 USD -0.7%
160 UTAH -0.7%
161 SCU -0.7%
162 CORN -0.6%
163 LOYO -0.6%
164 FORD -0.6%
165 LTU -0.5%
166 SFNY -0.5%
167 UCA -0.5%
168 CCU -0.5%
169 PUR -0.5%
170 BONA -0.4%
171 BUTL -0.3%
172 CLST -0.3%
173 MURR -0.3%
174 ALC -0.3%
175 IND -0.2%
176 PENN -0.2%
177 ARIZ -0.1%
178 WRIG 0.0%
179 OHST 0.0%
180 NCST 0.2%
181 UMKC 0.2%
182 ILL 0.3%
183 JAX 0.3%
184 WSU 0.3%
185 OU 0.4%
186 MIO 0.4%
187 WIU 0.4%
188 TAM 0.4%
189 FIU 0.5%
190 OSU 0.5%
191 MEM 0.5%
192 PRES 0.6%
193 KSU 0.7%
194 UTM 0.7%
195 KU 0.8%
196 ILST 0.9%
197 PAC 0.9%
198 MAN 1.0%
199 TNT 1.0%
200 UNCW 1.0%
201 GU 1.0%
202 HAMP 1.1%
203 CWM 1.1%
204 BELM 1.1%
205 LIUB 1.1%
206 UCLA 1.1%
207 SMU 1.2%
208 WISC 1.2%
209 FRES 1.2%
210 NW 1.4%
211 HOU 1.4%
212 CENT 1.5%
213 MVSU 1.5%
214 UNH 1.6%
215 CONN 1.6%
216 TUL 1.7%
217 UNCA 1.7%
218 TOWS 1.7%
219 USM 1.8%
220 DART 1.9%
221 SF 1.9%
222 FLA 1.9%
223 COLU 1.9%
224 MAR 1.9%
225 NEV 1.9%
226 ISU 2.0%
227 PROV 2.1%
228 NWSU 2.2%
229 BAY 2.2%
230 COLG 2.2%
231 DRKE 2.2%
232 WMU 2.2%
233 GSU 2.3%
234 GWB 2.3%
235 CMU 2.4%
236 COPP 2.4%
237 SFPA 2.4%
238 UALR 2.5%
239 CSF 2.5%
240 MINN 2.5%
241 WEB 2.5%
242 FAU 2.6%
243 IDST 2.7%
244 RUTG 2.7%
245 UNLV 2.8%
246 TTU 2.9%
247 NCC 2.9%
248 WAG 2.9%
249 HAW 2.9%
250 UK 3.0%
251 DREX 3.0%
252 TOL 3.1%
253 TU 3.1%
254 COLO 3.1%
255 UL 3.1%
256 UMD 3.2%
257 BRY 3.2%
258 POLY 3.3%
259 BRWN 3.3%
260 ULM 3.4%
261 GS 3.4%
262 MISS 3.5%
263 TXST 3.5%
264 UNO 3.5%
265 NU 3.6%
266 STAN 3.7%
267 UAPB 3.7%
268 CHAR 3.7%
269 JMU 3.7%
270 SEA 3.8%
271 SAMF 3.8%
272 PVAM 3.9%
273 TAMC 3.9%
274 FURM 3.9%
275 UAB 4.0%
276 TXSO 4.1%
277 DEL 4.1%
278 SHU 4.2%
279 RICE 4.3%
280 JSU 4.3%
281 EIU 4.4%
282 GRAM 4.4%
283 DEP 4.4%
284 TENN 4.4%
285 MSU 4.5%
286 ARST 4.5%
287 RICH 4.5%
288 ECU 4.6%
289 HOFS 4.6%
290 ND 4.7%
291 SAMH 4.8%
292 UTEP 4.9%
293 EWU 5.1%
294 EKU 5.1%
295 ALB 5.2%
296 NDST 5.2%
297 NTEX 5.2%
298 AKR 5.3%
299 LSU 5.3%
300 MSST 5.4%
301 SAC 5.5%
302 MIZZ 5.5%
303 ORSU 5.5%
304 IPFW 5.5%
305 AUB 5.5%
306 BUCK 5.5%
307 CCSU 5.6%
308 NDAK 5.7%
309 CREI 5.7%
310 HP 5.7%
311 APSU 5.8%
312 UTSA 5.8%
313 UGA 5.8%
314 HC 5.8%
315 SELU 5.9%
316 BU 5.9%
317 LMD 6.0%
318 ULL 6.4%
319 SH 6.6%
320 APST 6.8%
321 USA 6.8%
322 POST 6.8%
323 OHIO 6.8%
324 MCN 6.9%
325 LMU 7.0%
326 TSU 7.0%
327 MRSH 7.1%
328 RMU 7.3%
329 DAV 7.3%
330 ALAM 7.3%
331 JXST 7.7%
332 ALST 7.8%
333 BGSU 8.1%
334 TULS 8.5%
335 WINT 8.8%
336 ARK 8.9%
337 MIA 8.9%
338 CSB 9.0%
339 UCI 9.0%
340 CSN 9.3%
341 BRAD 9.6%
342 WCU 9.9%
343 SFA 10.0%
344 UMBC 10.3%
345 YALE 10.5%
346 HOW 11.4%
347 HARV 13.2%

BracketBuster matchups added

I've added BracketBusters matchups to the schedules. In the next few days look for an analysis of which teams are most likely to benefit.

Conference Tournament Simulations

As promised, I've started doing the conference tournament simluations in the separate pages. Basically I seed all of the teams at the end of each simulation and run a simulated tournament for each appropriate conference and then calculate RPI. For a more detailed description of what I do see here:

http://rpiforecast.blogspot.com/2008/12/conference-tournaments.html

To see your team's webpage, there are links near the tops of the overall page and individual team pages to those with conference tournaments included in the simulations.

Here's the overall page which includes conference tournament simulations in the RPI: http://www.rpiforecast.com/ct/index.html

And as an example, here's Duke's page with the tournament included: http://www.rpiforecast.com/teams/ct/Duke.html

SO the difference is the ct folders in the address. You should be able to find your team, however, by going to their page and clicking on the link.

Server Issues Fixed/Conference Tournaments

Well, the server issues seem to have been resolved, so hopefully no more problems going forward. As soon as I get a chance, I'm going to start adding (as I have done in past seasons) simulated conference tournaments for each of the 10,000 simulated seasons. The big hurdle in implementing that is figuring out all of the different formats and seeding schemes.

Server Issues

I'm having some serious issues with the hosting service I'm using to the point that It's virtually impossible to update the site. Hopefully that will be resolved soon. I had all of the data ready to update this morning, but was not able to do so. I did end up getting the main page and the conference pages up, but that took quite some doing. Some of the pages appear to be a few days old. That's because they somehow defaulted to a recent backup.