Examples are taken from S31 Full Analysis.
Standard formatting for a one-variable statistic set below: n = sample_size | min, q1, median, q3, max | mean +/- stdev | mode(s)
CRD – Cardboard
- Not picked up – The only interaction with the cardboard piece the whole game was landing on it any number of times, but all those times a dud/repeat was called. Exception: on Mystery, declining to risk even on a successful letter call also results in Not picked up.
- Bankrupted – The cardboard was picked up successfully, but was BANKRUPTed that same round. (no WC)
- Retired – The cardboard was picked up successfully, but the player did not solve the puzzle in the same round, losing said cardboard. (no WC)
- Claimed – The cardboard was picked up successfully, and the player solved the puzzle in the same round, claiming the prize. (no WC)
- Used – Wild Card only – denotes maingame use.
- Was bankrupt – A Mystery wedge was flipped and was, of course, BANKRUPT. If flipped and successful, it will fall under Bankrupted, Retired, or Claimed.
- Retired from risk – When a Mystery wedge is flipped, the other one loses cardboard status – it’s a static $1000 space the rest of the round. That’s all this denotes (and is technically redundant – adding up Bankrupted, Retired and Claimed counts for the other Mystery should give the same total).
- Bankrupted after claimed – For 1/2 Car tags, WC, and MDW, BANKRUPTing the cardboard after solving the puzzle to claim it (except WC – picking up the WC is instantly claiming it, since you don’t need to solve that puzzle to maintain possession of it in future rounds).
- Abandoned – A 1/2 Car tag was claimed and never BANKRUPTed afterwards, but was never paired with another 1/2 Car to complete the car.
- Completed – A 1/2 Car tag was part of a successful full Car win. Possible to have 3 halves in a single game with this result.
- Lost maingame – For WC/MDW, the player kept the wedge the whole rest of the game, but lost the maingame, leading to no real use of the cardboard.
- Went to BR – For WC/MDW, the player kept the wedge the whole rest of the game and won the maingame, leading to a 4th consonant or ONE MILLION for 100k.
SPN – All Spins
The 3rd column is the 2nd column divided by the number of rounds. For example, an excerpt of R1’s data:
R1:
Turquoise $2500 | 62 | 0.32 | 4.2%
BANKRUPT #1 | 48 | 0.25 | 3.3%
---------------------|------|------|-------
TOTAL | 1459 | 7.48 | 100.0%
Big Money wedges were hit 62 times, or 0.32 times (62/195) per R1. Note again that the divisor is the number of ROUNDS, not SHOWS: that is, for R5/R6, the divisor will be significantly less than 195, especially R6:
R6:
Free Play | 2 | 0.67 | 50.0%
Purple $650 | 2 | 0.67 | 50.0%
------------|---|------|-------
TOTAL | 4 | 1.33 | 100.0%
So in S31, there were 4 spins in R6, and an average of 1.33 spins per R6 – thus, 3 R6s in S31.
The 4th column is the 2nd column divided by the overall total. Thus above, each of Free Play and Purple $650 constituted half of all spins of R6 on the whole season.
Finally, all spins are combined into one chart, by each 1/24th on the Wheel. On such a big sample size, each 1/24th should have roughly an equal share of the pie, so to speak (low standard deviation).
MIL – MDW Hits
Simply specifying how many of the MDW hits were really LEFT/RIGHT BANKRUPTs, to clarify the SPN section.
BKT – Bankrupt trash
This is only pure cash, shown in the actual bank, lost to BANKRUPT. No gift tag, Mystery, Prize on Wheel etc. included.
FPL – Free Play Actions
Self-explanatory.
CL1 – Called letter stats #1
The 3rd column is the 2nd divided by the number of shows. So for example,
A | 585 | 3.00 | 7.4% Q | 3 | 0.02 | 0.0% TOTAL | 7884 | 40.43 | 100.0%
A was called 585 total times throughout every puzzle in the maingame of S31, or exactly 3.00 times per show on average, 7.4% of all letter calls. Q was meanwhile called only 3 times, for 0.02 times per show (or more accurately, once every 65 shows). There were an average of 40.43 letter calls in one maingame on average, which seems like a lot, but covers all duds & repeats, all vowel buys, as well as all speed-up calls.
CL2 – Called letter stats #2
The same data as CL1, but with repeats removed and split up by round. Is a letter significantly called less or more in a certain round? Here’s where to see.
CLF – Called letters’ frequencies
Now for CL1, take every called letter’s actual frequencies – that is, on a dud A, add 0 to A on the chart, on three T’s, add 3 to T’s on the chart – for essentially a massive weighted average. So in S31, we get this total line:
TOTAL | 13124 | 67.30 | 100.0%
The 3rd column is divided by 195 again, so in an average show, 67.30 letters were lit up on the puzzleboard in the maingame (of course, not including toss-ups).
FIL – First Letters
Self-explanatory. There should be few surprises in the maingame, the stat was mostly made for interesting BR insight.
RPT – Repeated letters
Self-explanatory.
PUD – Puzzle dissections
Self-explanatory – count how many times every letter shows up in a puzzle, for all puzzles in every round of a show set.
MEF – Maingame efficiency
A lot of data is presented concisely here (note: repeats can’t really be counted, since it could be on any frequency). Let’s look at the most popular maingame consonant, T:
T:
Dud | 110 / 150 | 73.3%
Single | 235 / 270 | 87.0%
Double | 223 / 232 | 96.1%
Triple | 119 / 119 | 100.0%
Quadruple | 38 / 41 | 92.7%
Quintuple | 14 / 14 | 100.0%
Sextuple | 1 / 1 | 100.0%
Septuple | 2 / 2 | 100.0%
Octuple | 2 / 2 | 100.0%
----------|-----------|-------
TOTAL | 744 / 831 | 89.5%
To explain:
- In 831 total R1-6 rounds in S31, T was called in 744 of them or 89.5% of all rounds.
- In 150 of those 831 puzzles, T was a dud. 110 of those times (73.3% of the time), that dud T was called. This is fairly inefficient – the dud T was only avoided 27.7% of the time.
- There were 4 T’s in a puzzle (Quadruple) 41 times. Those T’s went uncalled 3 of those times. And so on.
- Subtracting the Dud entry manually from the TOTAL, we get a fraction of 634/681. 681 times T was a good letter call, and it was called 634 of those times.
- Perfect efficiency for a letter would be 100% called on Single and above, 0% called on dud. Of course, T having high called percentages across the board is not surprising.
Obviously, more weight efficiency-wise should be put on the higher frequencies. In particular, P had a strong S31:
P:
Dud | 27 / 503 | 5.4%
Single | 89 / 250 | 35.6%
Double | 38 / 55 | 69.1%
Triple | 14 / 15 | 93.3%
Quadruple | 4 / 6 | 66.7%
Quintuple | 1 / 1 | 100.0%
Sextuple | 1 / 1 | 100.0%
----------|-----------|-------
TOTAL | 174 / 831 | 20.9%
BNR – Bonus Round
Self-explanatory: the BR called letter count total should be (number of shows)*4 + WC BR uses. The following tables by letter show an efficiency exactly alike to MFE above.
Finally, the non-dud total of each letter is summed up in one chart. The overall total is not too important, it’s really the average (for S31, 6.50 good letters per BR). You can see how in the extremes, there’s a range of 3 to 10 unique good letters in a BR, these extremes have been found to be consistent across seasons as well.
EFF – Efficiency By Frequency
Self-explanatory – a more direct look at efficiency. For each frequency 0, 1, 2… (Dud, Single, Double…), the percentage of times the actual letter is called. How low is the overall dud percentage, and how high are the rest?
TEW – Total Earned Winnings
Self-explanatory. 3rd column is divided by number of shows throughout, despite R5/R6 discrepancy.
BR | 2785150 | 14283 | 28.6% TOTAL | 9746854 | 49984 | 100.0%
So an average show produces just shy of 50k in S31, with 28.6% of those earnings coming from the BR.
MWS – Maingame Winnings Stats
Self-explanatory.
MSV – Missolves
Self-explanatory. For BR, some arbitrary judgment is used – in general, complete gibberish is ignored, partial guesses are ignored. Full guesses that make some type of sense are counted.
TLT – Total Lost Turns
Self-explanatory.
LTS – Lost Turn Stats
Self-explanatory.
TLB – Total Lost Banks
Like total bankrupt trash, cash only, for players who lose the round.
HM1, HM2 – Highest Multiples Left
For any one puzzle, take all the uncalled consonants in it. What is the highest frequency in the puzzle of all these uncalled consonants? Take all the letter(s) with this frequency, and those are your “highest multiple(s) left”. A measure of inefficiency. The letters themselves are only recorded for doubles or more.
HM2 simple sums up HM1 into two big charts.
ALC – All Letters Called (also Perfect Rounds)
Self-explanatory.
TRW – Total Round Winners
Self-explanatory, should be fairly random (uncorrelated to any one podium).
REL – Winner of Round Relative to Starter
E.g. if YELLOW starts a round, STARTER YELLOW, NEXT BLUE, and LAST RED.
BMF – Big Money Frequencies
Self-explanatory.
WFR – Waste frequencies
Self-explanatory, not as useful in S30+ but hey, less waste is better.
SSS – Spin Strength Statistics
Any huge outliers are probably due to editing, but in general the distribution should be accurate, hence giving the entire chart.
FSL, FSS – Final Spin Locations & Strength Statistics
With Pat these days, any spin beyond 72 is almost certainly due to editing (any black/white/Free Play is cut out for time). A couple <40 are probably also editing, but he’s been weakly spinning quite a bit too.
WCU – Wild Card Uses
Self-explanatory.
MYS – Mystery Stats
Self-explanatory.
EXP – Everything with Express
Explained with the stats themselves.
BUZ – Buzz-outs, Blanks
Self-explanatory.
PUI – Puzzle Info
Self-explanatory.
CAT – Categories
Self-explanatory.
PLA – Placing
Self-explanatory.
MAR – Margin of Victory
Difference between 1st and 2nd, 2nd and 3rd covered.
“Over minimum” considers a hypothetical absolute minimum game: round minimum on every maingame round (for the PP round, just the PP prize), with the toss-ups assumed solved. Hence, in a singles R4 game, 6k of toss-ups, 3 minimum rounds of 1k, 1 PP minimum round of nothing but PP –> “9k + PP” is the 0 point. R5 on doubles, 6k of toss-ups + 4 rounds of 2k, 1 PP round –> “14k + PP” is the 0 point. A negative amount would mean that toss-up(s) were triple stumped and the rest of the rounds could not make up for it – truly terrible, if it ever happens.
MC1 – Maingame Correlations #1
Main rounds: 0 | 0 / 0 | 1 | 11 / 31 | 35.5% 2 | 35 / 85 | 41.2% 3 | 25 / 65 | 38.5% 4 | 6 / 12 | 50.0% 5 | 1 / 2 | 50.0% 6 | 0 / 0 | ------|----------|------- TOTAL | 78 / 195 | 40.0% Average: 2.37 / 2.33 main rounds / winner
So in S31, 31 out of 195 maingame winners managed to win with just 1 main puzzle (not a toss-up) solved. Only 11 of those managed to win the BR as well. On average, BR winners solved 2.37 main puzzles on average, all maingame winners 2.33 main puzzles – nothing significantly different. Toss-ups have a more significant difference here.
In a big sample size like a full season, the more puzzles solved, the better the BR conversion rate should be – this signifies a dominating player.
PRZ – Prizes
Self-explanatory.
BRC – BR Correlations
How does winning the PP and sweeping toss-ups correlate to winning the maingame, and winning the BR? Also included is if the PP mattered (if the PP was absent / worth $0, contestant would have lost the maingame).
JKP – Jackpot
Self-explanatory. Included in S31+ for potential Jackpots for funsies, obviously subset of won Jackpots will be empty.
TRI – Trivia
Self-explanatory. Out of date in the past couple of years, but S29 did have 3 cases of it, and I’m prepared for the future if it ever returns…
BRA, BRW, BRT – Bonus Round “Pairs” by Amount, by Wedge, Total
Self-explanatory, numerator are solved BRs for each pair.
GEN – Gender
Self-explanatory.
woftracker 