Wisdom of Crowds - your predictions please!

[CusDamato]

that's why you can earn credits from good predictions

Thanks, didn't see that

[shortfuse122829]

Ok good, so I can still just guess for free? My dad was a gambler and he sucked at it lol, so yeah i don't really fuck w/ that tbh.

 

But I enjoy doing it just for fun. I don't guess on all the fights obviously, I have missed a couple events and really only guess on fights that I'm pretty familiar w/ the fighters but it's still kinda cool to just mess with

[MMATycoon]

yeah of course - we need people to do the predictions or we fall on our ass

[PBR]

another bad outing this time also

[shinault06]

I want a contract Terrance Dollar. Time to ko u ha

[MMATycoon]

Yep, not good that one.

 

It looks like we have a problem with people guessing, so I'm probably going to start weighting people who don't predict as many fights more highly and weighting down people who get fights wrong more.

[CusDamato]

I might be mistaken but it looks like the numbers from the bigfoot/mir card have changed to be a lot better then it was?

 

*EDIT, never mind i see you have been messing with it. I dont think the betting sites you linked with would like that though, it seems a little dodgy to change the numbers after the event.

[MMATycoon]

This is the third edit to the formulae, which I've always said would be an ongoing process. I'm not changing the raw data, just the interpretation of it, to see what gets the best results. Improving the formula by seeing how it predicts better the past, will mean better predictions in the future, which is what's important.

 

Not sure which betting sites you're talking about. I agree I need to make it clear what the deal is with redos when I try and sell these predictions.

 

I'm keeping track of the actual predictions in terms of credit bets anyway, on a tracker site, as of the last two events, so those will never get changed.

[CusDamato]

Not sure which betting sites you're talking about. I agree I need to make it clear what the deal is with redos when I try and sell these predictions.

 

I'm keeping track of the actual predictions in terms of credit bets anyway, on a tracker site, as of the last two events, so those will never get changed.

Ah i thought you had a post after the bigfoot/mir fight that said it was the first since you linked it to some betting sites, must of been the tracking thing. I thought people had payed for the tips then used them only to come back and see they are different a week later.

[MMATycoon]

No, nobody's paid anything yet. They won't be doing either unless we have better results than the last two events

 

This is the place we're tracking unit bets;

https://www.cappertek.com/picks.asp?shs=MMABetting.tips

[itakebrides]

Hey mike i just of one suggestion why don't you calculate what the percentage profit be of the best tip on a single card.what most pro bettor do is just bet one fight that there fairly sure on instead of every fight on card.I would love to see the data for the best tip on a card your nap as it were over a few fight cards and see how you do

[MMATycoon]

we have the units system in place now, which is basically that.

[PBR]

14th Mar UFC 185 - Pettis vs. Dos Anjos 12 1303 $-27.67 (-23%) $-51.36 (-43%) $-65.55 (-55%)

 

OUCH!!!!!!! -- worst one yet but i have a feeling most all did bad probably

[JLP]

Last night must have been rough too. If i read the page right, only a few guys picked Dos Anjos. I'm coming to the conclusion that this betting system works ok with mid card and lower card fights. Fighters that are well known, but are a rather large betting favorite. I dont know if its really conducive to actually making profitable bets long term because the consensus bets will always be rather low returns. High profile fights with close odds will likely not give us a consensus and the occasional upset like last night will make us look very bad.

 

I also dont really understand how to use unit betting here. If we cant adjust our units on a fight by fight basis then it isnt really doing anything. Last night for example, If I were actually betting like I used to I would have bet heavily on Theodoro and Jedrycysk and would not have placed a bet on any other fight on the card. I thought Pettis was going to win, but I wouldn't have bet on him because the ROI is no where near the associated risk that his odds give us.

What we are doing here is more of a pick em and really doesnt involve betting for profit from most of the managers.

[MMATycoon]

You aren't meant to look at it from a betting perspective - just pick who you think will win and don't take into account betting value at all. It's when combining everyone's perspective on who will win the fight where we decide if there's any value in placing a prediction.

 

It's been a rough few events but we have only had one fight of any significant betting value recently... Hopefully we'll get a better run of things soon, with a few more value bets. Even with 3 events in a row lost, that's still only a small overall loss compared to a lot of the single event profits we've had in the past.

[JLP]

You aren't meant to look at it from a betting perspective - just pick who you think will win and don't take into account betting value at all. It's when combining everyone's perspective on who will win the fight where we decide if there's any value in placing a prediction.

 

It's been a rough few events but we have only had one fight of any significant betting value recently... Hopefully we'll get a better run of things soon, with a few more value bets. Even with 3 events in a row lost, that's still only a small overall loss compared to a lot of the single event profits we've had in the past.

i interpreted the confidence slider as a betting value indicator. I was placing a much higher confidence level in fights that I would actually feel confident in betting on. What is the confidence slider for if thats not its intended use?

If we are just picking winners, I dont see how there will be enough variation between our picks and betting lines to convince someone to pay for our picks. We'll probably end up mirroring the betting lines for the most part. I guess if you see slight variation between our picks and a betting line, it could let you know if there is value. But I doubt that any serious bettor would trust our results over there own opinion once they realize how you are coming up with the picks.

[MMATycoon]

I believe it's actually good if we match the bookmaker odds sometimes / a lot of the time. That's showing that the equations are producing good, solid results.... Bookmakers often get it right, so if we match them, we're often getting it right too... However, it's where we differ that we're saying there's value.

 

The difference between the red bar and the orange bar is the value. A lot of the time this is significant.

 

As for confidence, this is how I describe it in the wiki.

"How confident are you in your prediction? You could also view this value as "predictability". If you don't think the fight is very predictable or you don't feel you know that much about these fighters then potentially it might be better if you don't predict that fight at all OR you can just give a low confidence value."

 

It shouldn't be anything to do with betting value (and neither should your picks in general) as you can leave that to us to calculate. So where you might not think there is much value in betting someone but still think it's a very high chance they will win, you should still make that prediction. As I say on the top of the WOC page, you shouldn't actually look at bookmaker odds before making your picks, as that will affect your decision making.

[Gonzasco]

Thanks for the VIP Prizes Mike. I'm ranked 27 for credits earned

[MMATycoon]

Good stuff. I would suggest for you personally to not predict as many fights. One bad event predicting all the fights like that and you'll soon lose all those credits earned.

[Guest]

When will we be able to cash out the credits? I have 70 days of VIP tgat I would like to spend on projects.

[ArtieBanks]

Man that last event the wisdom was not with the crowds, events like that are hard to predict because a lot of unknown fighters taking part.

[MMATycoon]

Yup, that was a real stinker. when it's so unknown I wish people just wouldn't pick. Since then I've done a bit of analysis and betting on favourites we're JUST in the positive then our bets on underdogs are at nearly 100% ROI. I don't consider us as having enough data yet but that's something I'll be keeping an eye on. I'm also entering a bit more data on how many high level fights each fighter has had at the time of the prediction, to see whether it correlates with profit / loss.

[MMATycoon]

When will we be able to cash out the credits? I have 70 days of VIP tgat I would like to spend on projects.

Not for a few more events - I wanna make sure everyone's happy with the equations (and I'm happy too obviously). It seems like maybe the amount given away is a bit low but then that could well be because we're on such a horrible run for the last 5 or 6 events.

[JLP]

I like the parlay WOC spit out today. Means is probably the biggest question mark.

 

Ronaldo Souza (against Chris Camozzi) Bookies: 93%, Us: 100%
Corey Anderson (against Gian Villante) Bookies: 78%, Us: 98%
Tim Means (against George Sullivan) Bookies: 61%, Us: 100%
Eddie Gordon (against Chris Dempsey) Bookies: 81%, Us: 94%

[MMATycoon]

Yup, would be nice if today went well Basically all on Machida though... I was just happy when he got bumped out the parlay because he was a lot higher than this until today.