With all this talk about tournament rankings once again, I decided it was time to actually have some data to back up opinions.
I wen't through the BOMB3 results from back in April. Not the results rather, but the ENTIRE bracket. Why? Simple, I wanted to use the ELO ranking system that Chess and a few other organazations use. Yet again, why? Because people have stated and thought in the past that this would be the best way to accurately rank Smasher's. Even I thought this, until I actually got around to working on it. (It is not the best way which I will now show).
My first point is time. It took me around 5 1/2 hours to go through the BOMB3 bracket that contained a little over 70 names. I had to individually calculate the result from each matchup. There were, in total, 279 matchups. Each player I gave a base rank of 1200. Then I entered in their number, the opponants number, and their score (1 for win, 0 for loss) into a Chess Calculator. I then had to reverse things to get the results for the opponant. This was not a fun process to say the least. It would be VERY unrealistic to think a database could be created by a single person, or even a group of people, to keep track of these things. The time alone it would take to run through a BOMB4/GS2/FC3 size bracket would be at least 10 hours for each one, or in GS2's case, closer to around 20 hours (lets see...120ish X 8 matches a person then double elim for top 22...thats alot of matches to calculate). Anyways, you get my point, no one has the time commitment, the savvy, the data, or the will to actually get the Chess Rankings working.
My second point is the accuracy I thought the Chess Ranking would produce did not meet my expectations.
Results from BOMB3 (real top 25)
1. Chu Dat (Ice climbers)
2. Dark Rain (falcon)
3. Husband (marth)
4. Wife (peach)
5. Chillin (fox)
5. Azen (link, yl, pikachu)
7. KM (marth, fox)
7. Mow (fox, mewtwo?)
9. Aho (marth)
9. Tejo
9. Tink
9. Inha (fox)
13. Ricky (link)
13. Steve W.
13. Oro (samus)
13. AlphaZealot (peach)
17. Gots
17. The Muffin King (mario)
17. Rickr (link?)
17. COMP (peach?)
17. Philly
17. Special K.
17. Mega
17. Mew2Matt (fox)
25. SJ
25. Meep
25. Rick V.
25. Ding
25. Shep
25. Mike P.
25. G-Regulate
25. ZorkFanDM
Results from Chess Ranking (top 25)
1: Chu Dat 1367
2: DarkRain 1326
3: Inha 1302
4: Mow 1296
5: Azen 1295
6: Wife 1280
7: Husband 1277
8: Tejo 1275
9: Steve W 1274
10: Chillin 1273
11: KM 1258
12: Tink 1255
13: Sleepy jack 1252
14: SS4Ricky 1251
15: Oro 1248
16: AlphaZealot 1232
17: Aho 1228
18: Special K. 1227
19: Mew2Matt 1224
20: Shep 1221
21: Philly 1209
22: Mega 1206
23: Muffin 1205
24: Comp 1204
25: RickR 1204
So, wheres the problem? Very simple, because many brackets run with bye because of double elim and an uneven number of players, some players will play way more matches than others, thus inflating their ranking to an unrealistic level (Inha/Mow for example). The same works for deflation, like the case of Husband and Chillin, they played less matches and as a result, their Chess Rankings did not accurately reflect their tournament performance.
My case in point: For everyone that wants to use a more complex ranking system than the one I proposed in the middle of the summer, good luck. You are simply being unrealistic. The data does not exist, the time commitment is too extravegant, and the results do not produce the accuracy that you think it would.
If you only want one reason: The data does not exist and will never exist in a form that can be easily compiled in a manner that could suite a more complex system. BOMB3 & BOMB4 are the only semi-large to large tournaments that I know of that still have brackets in their origanal form. I think the TG6 bracket was posted online awhile back, so I guess that makes three tournaments. If anyone has a hundred or so hours to spare, I guess you could get to work on it.