[Blax0r]

This may be a massively stupid question, but my current model uses data only from 250 or above events for betting only 250 or above events.

From your experiences, does historical performance from the challengers have any significant predictive power for the 250-and-above events?

Thanks in advance.

[Optional]

I think following the challengers can often give you an edge in ATP qualifying betting.

[Blax0r]

I think following the challengers can often give you an edge in ATP qualifying betting.

Hmm, and in turn, do you believe that ATP qualies data should be used for ATP tourneys?

I don't use qualies data either, which is probably a hole in my analysis.

[Optional]

For that tournament I take notice of it. You can often spot a decent 1st round underdog in good touch on the court surface coming up against a fave who might already be uncomfy on it.

I look at the previous quali results when assessing matchups, but don't do any mathematical type of capping, so you may be right in not including it.

[Blax0r]

For that tournament I take notice of it. You can often spot a decent 1st round underdog in good touch on the court surface coming up against a fave who might already be uncomfy on it.

I look at the previous quali results when assessing matchups, but don't do any mathematical type of capping, so you may be right in not including it.

I'm interested in hearing from people's experiences, which is exactly what you provided; thanks.

Still would be interested in hearing from others.

[EaglesPhan36]

I don't understand models in tennis betting. I have seen a lot of people talk about their models for tennis. WTF are these exactly? I remember Lou in particular routinely talking about this shit, the same guy who was betting large chalk in WTA quite frequently. If that is how a model works, fukk a model.

Not bashing you, I just have not seen anyone really explain how their model applies, what it looks at, etc. Dumb shits like me keep it simple, so I would enjoy hearing how yours works. Challengers are the same as any tournament, you get a sense of a player's form by seeing their results on that level. You have to look at who they play and how they do. Obviously someone rolling against players ranked in the 1000s isn't impressive, but someone winning Challenger titles week after week is someone to keep an eye on. I.E. Klizan this year.

[Blax0r]

I don't understand models in tennis betting. I have seen a lot of people talk about their models for tennis. WTF are these exactly? I remember Lou in particular routinely talking about this shit, the same guy who was betting large chalk in WTA quite frequently. If that is how a model works, fukk a model.

Not bashing you, I just have not seen anyone really explain how their model applies, what it looks at, etc. Dumb shits like me keep it simple, so I would enjoy hearing how yours works. Challengers are the same as any tournament, you get a sense of a player's form by seeing their results on that level. You have to look at who they play and how they do. Obviously someone rolling against players ranked in the 1000s isn't impressive, but someone winning Challenger titles week after week is someone to keep an eye on. I.E. Klizan this year.

That's now 2-0 for people who do pay attention to Challengers; I think I'll have to start scraping those statistics. Thanks for your input.

Also, I PM-ed you an overview of my model; I'm not sure if it's what everyone else does, but the methodology makes the most logical sense to me.

[silvap]

i've seen you talk a couple of times about the model (first noticed on the deep errani run on RG) and was curious about how it works - im assuming it's very math-based, which is on the far opposite end of my qualifications. I am relatively new to sports handicapping and mostly focus on soccer but tennis looks like it could be a profitable sport as well. having stated my lack of knowledge on the subject, i'd assume the more data and info you have on players should only help make your model more acurate - unless you are worried about extra time used to gather and analyze the extra data not making an impact on the results and therefore not being worth the effort. best of luck with your bets sir

[Blax0r]

i've seen you talk a couple of times about the model (first noticed on the deep errani run on RG) and was curious about how it works - im assuming it's very math-based, which is on the far opposite end of my qualifications. I am relatively new to sports handicapping and mostly focus on soccer but tennis looks like it could be a profitable sport as well. having stated my lack of knowledge on the subject, i'd assume the more data and info you have on players should only help make your model more acurate - unless you are worried about extra time used to gather and analyze the extra data not making an impact on the results and therefore not being worth the effort. best of luck with your bets sir

PM-ed you an overview of the model. Actually, I'm not manually grabbing the data, so gathering more information has a very small cost (boils down to about 10 extra key strokes).

Soccer is the sport I eventually want to be able to handicap, but the lack of statistics makes it extremely difficult. My guess is we'd have to count the # of scoring opportunities created (offense) or given (defense) per team?

Anyhow, tennis is my focus, and after reading my PM, will be possibly become a large focus for you later on haha.

Thank you for your thoughts.

[BrianLaverty]

I don't see how a math model can be too accurate in predicting tennis, a game which is so dependent on the mental aspect.

Models are much more successful in team sports, imo..... individual sports are just too psychologically driven.

[Blax0r]

I don't see how a math model can be too accurate in predicting tennis, a game which is so dependent on the mental aspect.

Models are much more successful in team sports, imo..... individual sports are just too psychologically driven.

I agree with you to an extent; the match outcome is completely dependent on one individual (rather than a team), so his/her mental/psychological state is magnified. However, I argue that this phenomenon should reflect in the data (higher volatility, significant deviation from mean performance under certain situations, etc.), which should be correctly interpreted in a model.

ex: chokers, how do we represent them? This would require point-by-point data, but I would derive an "importance" figure (pr(win match | win current pt) - pr(win match | lose current pt)) and look for a correlation between this statistic and the choker's deviation from his normal performance.*

In addition, team dynamics and coaching decisions seem a lot harder to model than a one-on-one sport. Personally, I would love to transition to basketball, soccer, or football, but it's a more complex game than tennis.

Anyhow, what do you think of the importance of Challenger tourneys? Do place a significant weight on a players performance in Challengers when analyzing an ATP 250-and-above bet?



*This isn't my idea; super tennis academics Klaassen and Magnus came up with this methodology when trying to see if points are iid. They aren't.

[BrianLaverty]

While I think you can overlook someone losing in a challenger tourney due to lack of interest, but if someone wins multiple challengers, they are more then likely both confident and in good form, and it will carry over well to the early round matches of the 250 tourney. Perfect example being Brian Baker.

[Optional]

While I think you can overlook someone losing in a challenger tourney due to lack of interest, but if someone wins multiple challengers, they are more then likely both confident and in good form, and it will carry over well to the early round matches of the 250 tourney. Perfect example being Brian Baker.

Lu, Soeda, Klizen ...