Of the Safari Zone (Or, curse you Chansey!): A strategy guide

[MrMasonMan]

OK lets say I threw a bait at one Chansey then threw 29 balls before it left. I now have one ball left and I find another Chansey what should I do? That's what I'm getting at. Say I have 10 balls or 5 balls left, in these cases which method is more likely to get me a Chansey? Surely the amount of balls remaining can effect your chances

[Cubes]

I edited the main post to reflect that.

[IzzetGuildmage]

OP is correct and this is why:

Notes: The chance of catching a Chansey within 3 balls is roughly the same using Balls only as it is using Bait and 4 balls. Once you start running out of Safari balls it is therefore recommended to only use Balls.

This is assuming the escape and catch chances in OP are accurate.

Also assuming bait always lasts 4 turns. It is 2-6, so taking the average should be accurate.

[IzzetGuildmage]

Yeah I was going to ask you what would be the better strategy based on how many balls you have left. I could imagine 1 ball, the best catch rate would be rock-ball?

Btw as soon as I started any method starting with bait I've caught like 0/20 Chanseys while with the ball ball ball method I caught like 3/7 (though it's probably just the RNG's fault)

Throwing a ball right away gives 4,8% chance of catching it.

Using a rock first gives 10% * 9,6% = 0,96% of catching it.

Stick to the ball.

[Reazh]

Chansey only appear in level 26 right? Because if you use a level 26 first pokemon team and max repel it could be easily to appear.

EDIT: Since I use max repels the rate of appearance of Chansey has greatly increased!

[IzzetGuildmage]

Chansey only appear in level 26 right? Because if you use a level 26 first pokemon team and max repel it could be easily to appear.

EDIT: Since I use max repels the rate of appearance of Chansey has greatly increased!

The only pokemon under 26 is Paras, with a 15% encounter rate. So you increase your odds of finding Chansey by 100/85*100-100=17,6%. Not too shabby for an extra 600 pokedollars.

However, you have to use a fast level 26, otherwise you risk being unable to run away from certain pokes (mainly Venomoth, but possibly Nidorino/a), wasting your time.

[SirBrian]

Bait once, spam balls.

[Reazh]

The only pokemon under 26 is Paras, with a 15% encounter rate. So you increase your odds of finding Chansey by 100/85*100-100=17,6%. Not too shabby for an extra 600 pokedollars.

However, you have to use a fast level 26, otherwise you risk being unable to run away from certain pokes (mainly Venomoth, but possibly Nidorino/a), wasting your time.

There are nidorans under level 26 and ryhorns level 25 or doesnt count?

I use krabby level 26 and it works perfectly.

[matchbow]

I'd like to see the math behind your calculations. You say that the chance of catching a baited pokemon with 4 balls is 9.6% but the actual chance is 9.26% so you either don't fully understand why your calculations work, or your calculations are wrong.

Upon doing my own calculations, it is clear that OP's calculations are wrong.

However, the conclusion he reaches is still correct, and the calculation is partly shown by IzzetGuildMage.

The anomaly here that may lead someone to believe that the mathematical conclusion disagrees with their empirical results is that the bait method has a higher catch rate only if you take into account battles that go longer than 15 turns. If you hit your baits on the exact perfect turns every time, these types of battles will occur 20% of the time.

Given the small sample sizes that we're inevitably dealing with because of Chansey's low encounter rate (typically 1-3 per safari run), it's not surprising that people might think that just throwing balls is a better strategy. It would take a huge number of safari runs to see a benefit to the bait strategy, especially given that the advantage is so small anyway (15% vs 10%)

[Cubes]

I'd like to see the math behind your calculations. You say that the chance of catching a baited pokemon with 4 balls is 9.6% but the actual chance is 9.26% so you either don't fully understand why your calculations work, or your calculations are wrong.

Where did I say that? Quote where you think I said it.

[matchbow]

Where did I say that? Quote where you think I said it.

So, you see that you get 4 bait balls every time you throw 1 rock ball, and when you add up the percentages, you get 9.8% rock vs. 9.6% bait

Excuse me? You said that if anyone was interested in your calculations, they should ask for them. That's what I'm doing here. To insinuate that I merely "think [you] said [something]" is a completely unprovoked insult

[Cubes]

I'm not insulting, I'm really fucking tired and couldn't see where you saw that. Sorry if you felt offended, I do appreciate your question.

I was just comparing percent chance of catch on a bait-induced ball vs. percent chance of catch on a rock-induced ball, without looking at escape rates or anything else. It was just for exposition, really, and that comparison and calculation doesn't even affect the outcome because rocks are REALLY fucking terrible.

Also, I didn't take into account throwing the baits exactly right, and in fact didn't account for throwing baits at all for the main part. I calculated first just rate of bait>ball>ball vs. just ball>ball>ball. Bait comes out better. Then I checked to see how having to throw another bait affects it. It still favors bait>ball>ball.

I've also already stated in this thread that the results are certainly more long-term than anything. However the bit about it taking 15 turns is wrong based on what I did.

Also, as I've stated before, (and I am trying to be a uguu about this), don't just say you have your own calculations without posting any work, proof, etc., then just say mine are wrong when I do have said work. If you're going to dispute it, you better bring your proof.

[matchbow]

The chance of catching with bait->ball->ball is 0%+2.14%+1.9%. The chance of catching with ball->ball->ball is 4.8%+2.64%+1.45%. It's all shown very nicely by IzzetGuildMage (regards), as I said.

I have literally no idea how you could have gotten the numbers you did for ball->ball.

Logically, you can reach my conclusion about baitless catching being better in this way.

On turn 1, the chance of catching without bait is 4.8%. The chance of catching with bait is 0%.

On turn 2, the chance of having caught it with bait-ball is :

P(getting to turn 2)*P(catching it) + P(having caught it on turn 1 already)

= 0.89*0.024 + 0

= 2.14%

(as shown above in the 2nd term of the bait->ball->ball probabilities)

The chance of having caught it with ball ball is:

P(getting to turn 2)*P(catching it) + P(having caught it on turn 1 already)

= 0.55*0.048 + 0.048

= 7.64%

Therefore on turn 1, naturally the chance of having caught it is better if you throw no bait. On turn 2, no bait is still better, but the margin by which it is better is shrinking. Therefore there is a point at which the bait method will surpass the ball method, and that point is around turn 15. Naturally the probability of catching it with both methods will approach 100% as the number of turns passed approaches infinity, but there is an absolute limit to the number of turns that can pass since you only get 30 balls, which leads us to the ~15% vs ~10% chance of catching it with each method, as shown by IzzetGuildMage.

I would like to point out that Cube4 still hasn't posted any calculations at all. I would also like to thank IzzetGuildMage for his nice chart

[Cubes]

I see what you're saying now. My calculations and method are actually the ones you use. But you're failing to take into account the probability of actually getting to the turns to throw the ball.

The chance of having caught it with ball ball is:

P(getting to turn 2)*P(catching it) + P(having caught it on turn 1 already)

= 0.55*0.048 + 0.048

= 7.64%

As a quick aside, I want to note that the normal catch rate isn't exactly double the bait catch rate (4.7 vs. 2.4) because of rounding and stuff. Izzet used 4.8 instead of 4.7. It shouldn't really affect things, but it's important to note.

What you want to say here is that there is a 55% chance to be able to throw 2 balls, with a catch probability of 7.29%. To reach that same probability with bait, you need

.89*.89*.89*.89*.89 = .5584, or, 6 turns- First turn bait, then 5 balls.

When you do this calculation, you get

2.4*.89 + 2.4*.89^2 + 2.4*.89^3 + 2.4*.89^4 + 2.4*.89^5 = 8.57%,

which is exactly what I put up on the first post.

Of course, this is assuming infinite baits. Baits last 2-6 turns. Assuming an average of a 4-turn bait, we get, for just balls

2.4*.89 + 2.4*.89^2 + 2.4*.89^3 + 2.4*.89^5 = 7.06%.

But, having 7 bait turns is .89^7 = 50%, and the calculation there becomes

2.4*.89 + 2.4*.89^2 + 2.4*.89^3 + 2.4*.89^5 + 2.4*.89^6 = 8.26%

So we have a 55% chance with just balls of catching chansey 7.29% of the time.

And we have a 50% chance with Bait and balls of catching chansey 8.26% of the time.

[skyscream]

so... much... math...

http://www.youtube.com/watch?v=HY-03vYYAjA

[heated]

This says the escape is calculated at the beginning of the turn.

If this is true, just throwing balls clearly outstrips bait/balling.

How do I know Bubblepedia didn't just copy it and get it wrong? Where are the sources?

Upon further consideration, the beginning of the turn may be defined as just after you choose your move.

Also, you're not including the math with being unable to predict when they stop eating and the turns on which they are not eating before you throw more bait.

IzzetGuildmage's chart also doesn't assume the turn bait is eaten counts as one of the bait turns. (his assumed average is thus 5 turns of eating)

Also, while Bubblepedia says that catch rate halves and resets when they stop eating, the previous link says that catch rate halves permanently but not to less than g=3. That would mean that for Chansey, (g=2) throwing bait does not decrease catch rate.

One thing that is relatively easy to test is the calculation of escape before or after you throw bait. Gather some data on the rate at which Chanseys flee right after the first bait OR right after a bait after a "Chancey is watching carefully" turn. If it is somewhere around 10%, escape is calculated after throwing stuff. If it is around 45%, escape is calculated before throwing stuff. Data significance for this should be at least 30 Chanseys.

Alternatively, Dratini with 35% and 5%. (at least 40 Dratini)

If Bubblepedia is correct about the catch factor, Izzet's chart would be modified to account for the 55% stay factor on each turn just before a bait.

If upokecenter is correct, the chart would still have to be modified for the 55% stay factor when not eating, but all catch chances would be the same as if not baiting. (making Bait the obvious choice)

If by some chance escape is calculated before your choice each turn, 55% stay factors would have to go on the bait turns, too.

[heated]

OP is correct and this is why:

Notes: The chance of catching a Chansey within 3 balls is roughly the same using Balls only as it is using Bait and 4 balls. Once you start running out of Safari balls it is therefore recommended to only use Balls.

This is assuming the escape and catch chances in OP are accurate.

Also assuming bait always lasts 4 turns. It is 2-6, so taking the average should be accurate.

With this adjusted for being unable to predict the exact endtime of eating, the chances are 11.61% and 10.67%, making baiting about 8.9% more effective for catching Chansey.

[Cubes]

Just went through that site you posted, and some things I've noticed.

"When the player throws bait, G is halved, but not to less than 3."

- If what the site says here is true, then there is no difference between normal balls and bait balls. Chansey's normal catch rate is 30 (via Bulbapedia). The "G" value in this website (I called it the safari-catch factor) for chansey would be

G = 30*100/1275 = 2.35294.

This is the value for G without any baits and balls. If we were to calculate with bait, we would halve that. EXCEPT G can't be less than 3. So the value for G with bait has to be 3, and so the numbers as is suggests that the catch rate is BETTER with bait than without. At the very least, 2.35294 would get rounded up and the catch rate for without bait would also be 3, meaning bait balls are just as good as regular balls. Now, I don't believe this, so I don't know how valid some of the other information on this website is.

"At the beginning of each turn (before commands are chosen), if a random number from 0 through 99 is less than 5 times X"

-The way this is worded leads me to believe that only the random number between 0-99 is chosen, and the calculation is done afterwards. I find it hard to believe that they would do the calculation before having thrown bait or rock, especially when it factors eating/angry. So to me, at least, it makes sense that only the number is determined before commands are chosen.

The site you posted is also sloppy in terms of labeling and defining variables. Although this site and bulbabedia do match up on most things, there's also nothing to suggest this site is any more accurate than bulbapedia, and perhaps less so, seeing as how these seem to be the archives to the site and the actual site is down. I've had this discussion before - we don't know for sure when values are calculated or when things like "eating" or "angry" end. This is the best guess we have.

[Elite]

I wish I could intelligently contribute to this discussion, but I didn't take statistics. (I wish I did).

However I can confirm that Reazh's method of using a level 26 Pokemon as a party lead + max repel does have a significant impact on Chansey encounters, as it negates all paras and most exeggcute (which is one of the most common encounters).

[Cubes]

However I can confirm that Reazh's method of using a level 26 Pokemon as a party lead + max repel does have a significant impact on Chansey encounters, as it negates all paras and most exeggcute (which is one of the most common encounters).

I wasn't sure how well this would work because of the steps limit. From experience in catching the roaming legendaries, limiting your options also limits how many encounters you get overall i.e. you were always going to run into x amount of chanseys in 600 steps, the only difference was whether you would have to deal with all the other crappy pokemon inbetween. But I haven't tested this foe Pokemmo safari zone, so it may actually help a lot. Some testing is in order, I think.

If you want to contribute to this, simply mark down both how many encounters you have total, and how many chanseys you encounter, while using repels. If the rate is higher than 4%, than you/Reahz may be on to something.

[heated]

The site you posted is also sloppy in terms of labeling and defining variables. Although this site and bulbabedia do match up on most things, there's also nothing to suggest this site is any more accurate than bulbapedia, and perhaps less so, seeing as how these seem to be the archives to the site and the actual site is down. I've had this discussion before - we don't know for sure when values are calculated or when things like "eating" or "angry" end. This is the best guess we have.

But where does Bulbapedia get its formulas?

[Cubes]

But where does Bulbapedia get its formulas?

It's a valid question, that can be asked of any site that doesn't show it's own testing. Bulbapedia and the site you listed both share similar enough information on the formelas, leading me to believe that the general formulas are correct. It's just that there were inconsistencies on the site you linked, whereas bulbapedia has none.

Your site gives out more information than bulbapedia, giving it more chances to be wrong. Bulbapedia says nothing about when forumlas are calculated and numbers are generated, or anything about the actual escape rates for the pokemon.

Essentially, your site says everything bulbapedia says, but also says a little bit more. The problem is in judging whether that little bit more is also correct. The inconsistency I showed in an earlier post hurts the credibility of the rest of the information, because if that site allowed one thing to be wrong, it could just as easily put up other things that are wrong too. Do you see what I'm saying?

There's the base information, which bulbapedia and upokecenter have largely the same, and then there's extra information, of which bulbapedia does not make any statements, while upokecenter does say something about, however at least one of their extra statements has been proven incredibly inconsistent.

[heated]

Yeah I just want to know where all of this info comes from.

Also can you edit your first post to include total catch chances with Bait/Ball and Ball/Ball?

And can you include the Super Repels in hotbar with a lvl 26 pokemon at the head of your team to decrease pokemon other than Chansey by about 20%?

[Cubes]

Yeah I just want to know where all of this info comes from.

Also can you edit your first post to include total catch chances with Bait/Ball and Ball/Ball?

And can you include the Super Repels in hotbar with a lvl 26 pokemon at the head of your team to decrease pokemon other than Chansey by about 20%?

-Yeah, I would like that too.

-Which percentages exactly? I could put that you'll catch on average 1 in 13 chanseys using baits and 1 in 14.5 chanseys using just balls, if you want.

-If I get more evidence that this is true I'll edit the main post. You can contribute by doing what I told Elite to do. I'll credit anyone who provides data.

[heated]

-Which percentages exactly? I could put that you'll catch on average 1 in 13 chanseys using baits and 1 in 14.5 chanseys using just balls, if you want.

-If I get more evidence that this is true I'll edit the main post. You can contribute by doing what I told Elite to do. I'll credit anyone who provides data.

"With this adjusted for being unable to predict the exact endtime of eating, the chances are 11.61% and 10.67%, making baiting about 8.9% more effective for catching Chansey."

(using basic probability and the catch rates using bait/balls according to Bulba)

http://bulbapedia.bu...dix:Repel_trick

I'm pretty sad that repel doesn't work for fishing.