MAIN FORUM

  • someclevername

    UFC Fight Night 160 / UFC on ESPN+ 18

    Saturday 9.28.19 – Royal Arena – Copenhagen, Denmark

    ESPN+ Main Card – 1pm CDT

    Jack Hermansson (185) vs Jared Cannonier (185)
    Mark Madsen (155) vs Danilo Belluardo (156)
    Gunnar Nelson (171) vs Gilbert Burns (171)
    Ion Cutelaba (205) vs Khalil Rountree Jr. (205)
    Michal Oleksiejczuk (205) vs Ovince Saint Preux (206)
    Nicolas Dalby (171) vs Alex Oliveira (171)

    ESPN+ Prelims – 10am CDT

    Alen Amedovski (185) vs John Phillips (186)
    Alessio Di Chirico (186) vs Makhmud Muradov (185)
    Siyar Bahadurzada (171) vs Ismail Naurdiev (171)
    Brandon Davis (146) vs Giga Chikadze (146)
    Macy Chiasson (136) vs Lina Lansberg (136)
    Marc Diakiese (155) vs Lando Vannata (155)
    Jack Shore (135) vs Nohelin Hernandez (136)

  • gje627

    Per before … and I don’t want to get into the math, but it’s fun … according to my numbers … below is the best lineup construction for tomorrow ….

    Importantly, this is for those who MME and want to “cover” both sides of the fight ….

    1) Select one (1) fighter from each of the three (3) fights below for a total three (3) total lineup spots:

    — Cannonier/Hermansson
    — Cutelaba/Rountree Jr.
    — Amedovski/Phillips

    2) Select two (2) fighters from the three (3) fights below for a total of two (2) lineup spots:

    — Belluardo/Madsen
    — Oleksiejczuk/Saint Preux
    — Hernandez/Shore

    3) Select the one (1) best fighter from any of the remaining fights:

    — Burns/Nelson
    — Dalby/Oliveira
    — Di Chirico/Muradov
    — Bahadurzada/Naurdiev
    — Chiasson/Lansberg
    — Diakiese/Vannata
    — Chikadze/Davis
    _____

    By using this formula, you can “cover” both fighters in each of the first five (5) positions (with the sixth (6th) and final position of a lineup being the wildcard (filled again, with the best remaining fighter that fits within the salary cap).

    Based on the match, if you do this, there is a 70.42% probability that you will have at least one (1) 6/6 lineup in only 21 lineups built and a 52.04% probability that one of the 6/6 lineups will be optimal (best possible)

    ———

    Also worth noting?.

    Select any six (6) fights …. Cover both sides of each fight ….

    84.63% probability that you will have a 6/6 lineup within the first 24 best lineups built. The percentage probability of your 6/6 will depend on which six (6) fights you choose to include.

    FYI: Salary cap constraints included in all numbers above ….

    EDIT: To put the above in context …. there are 180,844 possible lineups at or under the $50,000 salary cap for the 13 fights tomorrow. Of these, there are 82,958 possible lineups that do not include any fight stacks. Of these 82,958 lineups, the highest probability of any one lineup being 6/6 equals 3.15%. …. Meaning, combinatorial optimization (i.e. not just selecting the best individual fighters, but instead the best fighters that fit best together in a 6-fighter lineup) is extremely important ….

    ——-

  • gje627

    Dang ….

    I speak in numbers ….

    But I just read everyone else’s thoughts in this thread now (i.e. after I posted my numbers above) …. so I +1-ed pretty much every post before mine ….

    So, as far as I can tell, everyone in this thread aligns exactly according to all my “numbers” !!!

    Smart people here ….

    Though I will say, I’m talking in lineups and not just individual fighters ….

    So maybe (?) I added a little something ….

  • gje627

    @someclevername said...

    You guys see DK is testing out Captain mode for mma? I got an email yesterday – calling it Rambo lol

    *Did they run a contest like this last week and I missed it?

    I hate showdown mode …. Seems like DK is putting all real-money contests into showdown, regardless of sport, and totally diminishing their regular (skill) offerings.

    Showdown in MMA …. Captain spot will be a total crap-shoot …. Once again, as is the most significant problem with showdown contests, only MME players will even have a remote chance of winning ….

    BTW … I’m an MME player, I generally max-enter, but trying to get more refined and get away from that with fewer entries ….

  • someclevername

    @gje627 said...

    So maybe (?) I added a little something ….

    You always add a lot, gje!!

    I just like watching the fights and usually do the ol’ single bullet, but I always find your information interesting (when it isn’t way over my head lol). I know you help out a lot of folks and it’s pretty cool that you share.

  • someclevername

    @gje627 said...

    Showdown in MMA …. Captain spot will be a total crap-shoot

    I feel this is going to help me find out exactly how much patience I have with bad mma beats.

  • callmetouch

    I appreciate the help in letting me know u can watch previous cards but can someone confirm I will be able to watch the full event 5-6 hours after it has ended? If so I will avoid knowing results and watch them once I am available. On espn plus of course

  • gje627

    @someclevername said...

    I feel this is going to help me find out exactly how much patience I have with bad mma beats.

    Guaranteed … regardless of my “can’t miss Captain” … I will always be on the wrong side of …

    A) No Contests
    B) Draws
    C) DQs

    See last card …. We almost had all of the above (no DQs) … but a very late cut that stopped a fight where the underdog (who I was on … Ariane Carnelossi) lost by “TKO” with only 2 minutes left in the final round because of said cut … in a fight she (my dog) very well could have / should have won.

    With that said, glad to see that Stephens/Rodriguez is back-on in only 3 weeks ….

    Based on Yair’s response to his illegal eye-poke … I hope he gets smoked in Boston by Jeremy !!!

  • harryf24

    Yes you can access the fights on espn+ after they are over. I actually enjoy this way to watch the fights more as you can fast forward in between the fights.

  • someclevername

    @gje627 said...

    A) No Contests
    B) Draws
    C) DQs

    D) Toe to the eye
    E) Surviving a choke til the bell and the ref calling it after

    …not exactly in the same category but still in the “wtf, really?” family lol.

    Strongly agree with Jeremy smoking Yair. Thought he was going to do it the first time and now he has extra incentive.

  • someclevername

    @harryf24 said...

    Yes you can access the fights on espn+ after they are over. I actually enjoy this way to watch the fights more as you can fast forward in between the fights.

    I should just fast forward through the fights I picked lol.

  • gje627

    @someclevername said...

    Strongly agree with Jeremy smoking Yair. Thought he was going to do it the first time and now he has extra incentive.

    Speaking of which … did you see the co-main in Boston?

    Great fights !!!

    Donald Trump / Nancy Pelosi
    Ivanka Trump / Any Old Rock (as in “dumb as a rock”)

    100% Pelosi here !!!

    Unsure about Ivanka and Any Old Rock … could be a draw …. :)

    Maybe Ivanka / AOC (Alexandria Ocasio-Cortez)?

    Sorry, as usual … I get “way-off” topic.

  • Driven2ThEdge

    @gje627 said...

    By using this formula, you can “cover” both fighters in each of the first five (5) positions (with the sixth (6th) and final position of a lineup being the wildcard (filled again, with the best remaining fighter that fits within the salary cap).

    Based on the match, if you do this, there is a 70.42% probability that you will have at least one (1) 6/6 lineup in only 21 lineups built and a 52.04% probability that one of the 6/6 lineups will be optimal (best possible)

    Thanks for the write-up. I have a few questions and I feel like your math or assumptions might be a bit off, but perhaps I’m misunderstanding.

    First, maybe I misunderstood your 3 grouping scheme, but I understand that it should yield every combination of all 3 fights from group 1, every combination of 2 out of 3 fights in group 2, and then whatever is best left over from group 3. But then you say that you build 21 lineups with this? But from my count, there are 8 combinations from group 1, and 12 combinations from group 2, yielding 96 total lineups if you want full coverage as you have laid it out.

    Secondly, can you explain there the following number comes from “70.42% probability that you will have at least one (1) 6/6 lineup”? It seems like it probably comes from pairing the best odds fighter from group 3, but you will often be left with not so much money left over, so you might often have to pick an underdog from group 3.

    But my main concern is with the following: “21 lineups built and a 52.04% probability that one of the 6/6 lineups will be optimal (best possible)”. SURELY, you don’t mean that one of your 21 lineups you created will be optimal, i.e., guaranteed to win the GPP. But I’m not sure what you actually mean. Thanks, and feel free to get as complicated with the math as you want in the response.

  • thevogaman

    I was thinking that the Mark O Madness ownership would be sub 40%. Hermansson will be highest owned. Definitely seeing the Jack Shore buzz also.

  • gje627

    @Driven2ThEdge said...

    Thanks for the write-up. I have a few questions and I feel like your math or assumptions might be a bit off, but perhaps I’m misunderstanding.

    First, maybe I misunderstood your 3 grouping scheme, but I understand that it should yield every combination of all 3 fights from group 1, every combination of 2 out of 3 fights in group 2, and then whatever is best left over from group 3. But then you say that you build 21 lineups with this? But from my count, there are 8 combinations from group 1, and 12 combinations from group 2, yielding 96 total lineups if you want full coverage as you have laid it out.

    Secondly, can you explain there the following number comes from “70.42% probability that you will have at least one (1) 6/6 lineup”? It seems like it probably comes from pairing the best odds fighter from group 3, but you will often be left with not so much money left over, so you might often have to pick an underdog from group 3.

    But my main concern is with the following: “21 lineups built and a 52.04% probability that one of the 6/6 lineups will be optimal (best possible)”. SURELY, you don’t mean that one of your 21 lineups you created will be optimal, i.e., guaranteed to win the GPP. But I’m not sure what you actually mean. Thanks, and feel free to get as complicated with the math as you want in the response.

    Thanks for the reply …. I completely understand what you’re saying … and your questions are both ….

    an accurate interpretation of what I said … and 100% spot-on ….

    Absent my “typical long-winded” explanation ….

    Please see my very first sentence in said post …. That being ….

    “Per before … and I don’t want to get into the math, but it’s fun … according to my numbers … below is the best lineup construction for tomorrow ….

    Note the according to my numbers part ….

    In other words … this is NOT strictly combinatorics as in the “all combinations sense” BUT instead “combinatorial optimization”

    i.e. we disregard combinations of events that are so unlikely to occur together that they can, for all intents and purposes, be totally disregarded

    E.g. today we will have a snowstorm and a tornado AND today we will set an all-time record high temp and an all-time record low temp AND today we will have 20 inches of rain and widespread wildfires due to dry brush ….

    Okay weird example, but hope you get my point ….

    In other words, quite a few combinations are completely disregarded based on a number of factors …. I can explain the formula/algorithm if you really want … probably would help … and make more sense … but would also take forever ….

    But … for example … a lineup like:

    Danilo Belluardo, Lina Lansberg, Ovince Saint Preux, Siyar Bahadurzada, Nohelin Hernandez, Nicolas Dalby ….

    is completely disregarded due to its very, very, very low likelihood of being 6/6 … let alone optimal ….

    Also note, none of “within so many lineups” numbers are a 100% sure thing ….

    Regarding the probability numbers … there are a lot of factors here too (another formula/algo) but suffice to say here …. only a very small portion of the formula includes anything more advanced than the basic rules of probabilities … including the addition rules, the multiplication rules, both for dependent and independent events, as well a conditional probability rules, both for dependent and simultaneous events) … from there it’s simply applying these rules (where they apply in the formula) in a combinatorial fashion (pretty much said in previous sentences, bu thought I should add) … to include a measure of error.

    ….

    Nonetheless, I have tested this on historical data …. and if you can pick the right fights to include you can almost always achieve 6/6 within 20 lineups, give or take a few ….

    One thing worth mentioning (emphasizing) you really need to pick the right fights even more precisely to be optimal …. To explain …

    There are 1,716 possible “fight combinations” (not fighter) including no stacks and at or under the $50K salary cap …. “Covering the Top 20” for each fight combination would give you 34,320 total lineups and that is only if you did 6-6 combinations …. i.e. you will absolutely not be able enter each and every one …. again the need for the “optimization” part ….

    EDIT: Sorry … a lot of edits ….

    EDIT 2: Not sure if this helps … a math explanation would be much easier, but it would take forever to explain …. Not difficult to explain just a lot-o’, lot-o’ words …. probably at least 3 pages worth (or more) in the forum….

  • someclevername

    @thevogaman said...

    I was thinking that the Mark O Madness ownership would be sub 40%. Hermansson will be highest owned. Definitely seeing the Jack Shore buzz also.

    Where do you think Austrian Wonderboy comes in? He has to come in low, right? Burning a bunch of people last time out vs Chance (thanks Chance!) and now being 9k vs a usually very game Siyar (oooh that liver kick to Chagas was nasty – same card as the Merab choke robbery oddly enough). I have thoughts about taking Naurdiev at low ownership. Crazy thoughts. Help me.

  • thevogaman

    @someclevername said...

    Where do you think Austrian Wonderboy comes in?

    I’m thinking he’ll be sub 20%. I’d guess that Davis, Muradov, and Diakiese will all go lower owned in that same range too.

    My only concern with Austrian Wonderboy is the matchup leads me to believe it goes to decision.

  • gje627

    Agree ….

    Using only 5Dimes odds (and nothing else) converted to remove vig, and expressed in percentages ….

    Win: 62.79%
    ITD: 23.36%
    Win Difference between Opponent: 25.59%
    ITD: Difference than opponent: 1.95% (only … very small)

    Salary: $9000 (Siyar $7200)

    5th highest salary, 5th highest win difference, 15th highest ITD difference …. Neither Sub (9.87%) or TKO/KO (16.93%) establish a clear road to early victory.

    Note about the props … ITD is a prop as are sub and TKO/KO …. All three are separate props so you can’t add Sub + TKO/KO to equal “exactly” ITD.

  • gje627

    @someclevername said...

    I have thoughts about taking Naurdiev at low ownership. Crazy thoughts. Help me.

    My thoughts?

    Not crazy … but really depends on your other 5 fighters too ….

    I’ll probably play 40-50 lineups tomorrow and I’ll probably be 0% or maybe include one flyer ….

  • someclevername

    @thevogaman said...

    My only concern with Austrian Wonderboy is the matchup leads me to believe it goes to decision.

    Yeah, I tend to agree. This sounds dumb as hell, but I didn’t like Siyar rolling into ceremonials donning the hoodie & sunglasses. Hate that – can’t see what I want to see. I must now create my own narrative lol.

  • someclevername

    @gje627 said...

    Maybe Ivanka / AOC (Alexandria Ocasio-Cortez)?

    Politics aside – would take AOC all day purely based on her bartending experience. My sister is a bartender (a.k.a. psychiatrist, babysitter, bouncer, marriage/relationship counselor) and it takes a great deal of mental strength to deal with drunks. Also physical strength (kegs are fun!).

  • someclevername

    @gje627 said...

    I’ll probably play 40-50 lineups tomorrow and I’ll probably be 0% or maybe include one flyer …

    Well, that doesn’t sound great lol

  • gje627

    @someclevername said...

    Well, that doesn’t sound great lol

    Ended-up with 42 lineups … one Austrian Wonderboy (2%).

  • Driven2ThEdge

    @gje627 said...

    Note the according to my numbers part ….

    In other words … this is NOT strictly combinatorics as in the “all combinations sense” BUT instead “combinatorial optimization” …

    Thanks for the response, and yes, I understand. Everything you say is completely reasonable, except the figure of having a 52% chance of having the optimal lineup in your top 21 lineups. I urge you to take a closer look at your methodology regarding this, because I strongly believe it’s off by an order of magnitude. And I completely understand that this is with a mix of 5dimes odds with some points derivation from outcomes, and that not all outcomes are equally likely.

    ———————

    Just so you don’t think I’m a crank, I’ll provide some information so you know I’m not full of it (also I have a PhD in a field heavily related to probability and statistics). I figure you’re doing something along the following lines:
    1. Scrape odds from bestfightodds.com (or copy/paste odds from 5dimes into excel, do some conversions, and export to csv)
    2. Convert the odds to percentages and normalize to get rid of the vig.
    3. From this point it’s fairly easy to get possible combinations of lineups with the highest odds of being 6/6. This is just multiplying the individual probabilities (or preferably adding the logs of the probabilities). Since the total number of combinations is enormous, you’ll want to just search through the valid ones first (salary wise).
    4. Alternatively, if you were just interested in the single highest probability of a 6/6 lineup, you would use a technique called linear programming. (This particular lineup always has A LOT of lineups on the big draftkings tournaments).
    5. Here’s where it gets more challenging — when you want to assign DK points. You probably either assign points for specific outcomes (like win by KO/TKO), or perhaps you have a probability distribution per outcome (perhaps also conditioning on the weight class).
    6. If you go the assignment of points method, then you simply have a larger optimization problem.
    7. If you go the distribution method, you now need to turn to monte carlo methods. If you go this route, I would recommend simulating the entire event many times (say 100000) and for each “event” use linear programming to determine the optimal lineup for that simulated event. You would do this by first probabilistically determining the winners each time, then sampling from the distribution, then setting each fighter to be a binary variable in the LP objective function. You would also have a salary inequality constraint and an equality constraint to guarantee exactly 6 fighters chosen.

    Anyway, there are 2 reasons I know that something is off about either your code or methodology. 1) I just ran my simulations and the most common optimal (edit: 21) lineups account for only 1.4% of all events. 2) You would be filthy filthy filthy rich if you could hit the nuts half the time. It would cost only $315 to entire 21 lineups, for a half chance to win $25k…

  • callmetouch

    All these simulations are cute and all but I’m editing 150 $1 lineups by hand on paper. Back in my days we didn’t have simulations

  • someclevername

    @callmetouch said...

    All these simulations are cute and all but I’m editing 150 $1 lineups by hand on paper. Back in my days we didn’t have simulations

    Wide ruled or college?

  • X Unread Thread
  • X Thread with New Replies*
  • *Jumps to your first unread reply

RotoGrinders.com is the home of the daily fantasy sports community. Our content, rankings, member blogs, promotions and forum discussion all cater to the players that like to create a new fantasy team every day of the week.

If you or someone you know has a gambling problem, crisis counseling and referral services can be accessed by calling 1-800-GAMBLER (1-800-426-2537) (IL). Gambling problem? Call 1-800-GAMBLER (NJ/WV/PA/MI), 1-800-9-WITH-IT (IN), 1-800-522-4700 (CO), 1-800-BETS OFF (IA), 1-888-532-3500 (VA) or call/text TN REDLINE 1-800-889-9789 (TN).